Keywords

1 Introduction to the Cellular Functions of Peroxisomes

Peroxisomes are single membrane-bound organelles that can be found in all eukaryotic cells with the exception of spermatozoa and mature erythrocytes (Novikoff et al. 1973). Peroxisomes display a high variability in their enzyme content and metabolic tasks that can be adjusted according to cellular needs. The enzymes in the lumen of peroxisomes are often highly concentrated and tightly packed to form crystalline inclusions that are visible as electron-dense structures. The beta-oxidation of fatty acids and the detoxification of the hydrogen peroxide are regarded as the central and most conserved functions of peroxisomes (Cooper and Beevers 1969; Lazarow and DeDuve 1976). Furthermore, the beta-oxidation pathway is linked to the synthesis of signaling molecules, like phytohormones in plants (Baker et al. 2006; Kienow et al. 2008) and pheromones in Caenorhabditis elegans and insects (Joo et al. 2010; Spiegel et al. 2011). Mammalian peroxisomes have a key function in the biosynthesis of ether lipids and bile acids (Wanders and Waterham 2006a). Peroxisomes house important steps of penicillin biosynthesis in some filamentous fungi (Meijer et al. 2010; Müller et al. 1991), but also certain enzymes required for the biosynthesis of Vitamin K1 in plants (Widhalm et al. 2012) or the synthesis of siderophores required for iron uptake and virulence of Aspergillus species (Gründlinger et al. 2013). Depending on the metabolic state of the cell, the number of peroxisomes can be dynamically regulated either by the proliferation of peroxisomes or their selective autophagic degradation via pexophagy (Grunau et al. 2011; Opaliński et al. 2011; Till et al. 2012; Tower et al. 2011).

Defects in peroxisome function are the molecular cause for human inborn errors that are caused by mutation of single metabolic enzymes (Wanders and Waterham 2006b) or genes coding for proteins that are required for the biogenesis of the organelles (Steinberg et al. 2006). The peroxisomal biogenesis disorders (PBDs) form a spectrum of autosomal recessive metabolic disorders that are collectively characterized by abnormal peroxisome assembly and result in multisystemic disorders that often lead to death in early infancy (Baes and Van Veldhoven 2012; Nagotu et al. 2012; Waterham and Ebberink 2012). Furthermore, the physiological function of peroxisomes contributes to the cellular protection mechanism against the progressive brain damage and cognitive decline caused by Alzheimer’s disease (Fanelli et al. 2013; Kou et al. 2011; Lizard et al. 2012).

The formation of peroxisomes depends on specific biogenesis factors, the peroxins (Distel et al. 1996). To date, 34 different peroxins have been described. In general, they are involved in the six key stages of peroxisomal biogenesis which comprise the (1) de novo formation and (2) proliferation of peroxisomes, (3) their inheritance and (4) regulated degradation by an authophagic process called pexophagy as well as the import of (5) peroxisomal membrane and (6) matrix proteins (Fagarasanu et al. 2010; Islinger et al. 2012; Liu et al. 2012; Platta and Erdmann 2007b; Theodoulou et al. 2013; Hasan et al. 2013).

In this review, we will discuss the peroxisomal matrix protein import with emphasis on the function of the peroxisomal membrane complexes that are involved in the ubiquitination and energy-consuming dislocation of the dynamic import receptors and describe their concerted function as receptor export machinery, the peroxisomal exportomer (Platta et al. 2013).

2 Peroxisomal Matrix Proteins Are Imported by Cycling Receptors

All peroxisomal proteins are encoded in the nucleus, synthesized on free ribosomes in the cytosol and imported posttranslationally. Most remarkably, peroxisomes are capable to accommodate fully folded proteins. Frequently, even oligomeric or cofactor-bound proteins are imported (Leon et al. 2006a; Girzalsky et al. 2009). The import of folded proteins distinguishes peroxisomes from other organelles like chloroplasts or mitochondria, which all import unfolded proteins, but it makes them comparable to the Tat (Twin-arginine translocation) pathways of bacteria and thylakoid membranes (Albiniak et al. 2012; Palmer and Berks 2012). However, in contrast to these translocation systems, peroxisomal matrix protein import is facilitated by dynamic receptors that cycle between a soluble state in the cytosol and a membrane-associated state at the peroxisomal membrane (Hasan et al. 2013; Liu et al. 2012; Platta and Erdmann 2007a). The import cycle can conceptually be divided into five steps, comprising (1) cargo recognition in the cytosol, (2) receptor–cargo docking at the peroxisome, (3) cargo translocation across the membrane, (4) cargo release into the matrix, and (5) receptor ubiquitination and export back to the cytosol (Fig. 15.1).

Fig. 15.1
figure 1

Peroxisomal PTS1-protein import in Saccharomyces cerevisiae. Peroxisomal matrix proteins are imported into peroxisomes via soluble receptors, which shuttle between the cytosol and the peroxisomal membrane. The matrix proteins are recognized by their peroxisomal targeting signal 1 (PTS1) in the cytosol via the receptor Pex5p, forming a receptor–cargo complex. At the peroxisomal membrane this complex binds to the docking complex (Pex13p, Pex14p and Pex17p), and this binding leads to the formation of a transient pore, whose exact molecular constitution is still under discussion but at least contains Pex5p and Pex14p. In the next step, the cargo is translocated into the peroxisomal lumen in an unknown manner; Pex8p might be involved in the receptor-cargo dissociation. At the end of the receptor cycle, the receptor is removed from the membrane and transported back to the cytosol for another round of import. To this end, Pex5p is monoubiquitinated by the ubiquitin-conjugating Pex4p (E2), which is anchored at the membrane by Pex22p, and by the ubiquitin ligase Pex12p (E3), which forms the RING-finger complex together with the other ubiquitin ligases Pex2p and Pex10p. The ubiquitin signal leads to an ATP-dependent dislocation of Pex5p from the peroxisomal membrane, performed by the Pex15p-anchored AAA peroxins Pex1p and Pex6p. Prior to a new round of import, the ubiquitin moiety is removed from the receptor

Full size image

The events concerning the cargo transport from the cytosol to the peroxisomal lumen are discussed in detail in Chaps. 13 (Effelsberg et al.) and 14 (Bartel et al.) of this book. In brief, proteins destined for the peroxisomal matrix usually harbor a peroxisomal targeting sequence (PTS). Most peroxisomal matrix proteins carry a C-terminal PTS1-sequence, which is recognized by the PTS1-receptor Pex5p, while a subset of proteins displays an N-terminal PTS2-sequence via which they are ferried to the peroxisomal membrane by the PTS2-receptor Pex7p and its co-receptors, like S. cerevisiae Pex18p or P. pastoris Pex20p. Here, the cargo-bound PTS-receptors interact with constituents of the docking complex (Pex13p, Pex14p), which results in the formation of a transient import pore and finally the transloction and release of the cargo.

Subsequent to the liberation of the cargo, the PTS-receptors return to the cytosol for further rounds of matrix protein import (Fig. 15.2). This retrotranslocation is facilitated by the peroxisomal receptor export machinery, the exportomer (Platta et al. 2013). The monoubiquitination of the PTS-receptors is regarded as a central event in this process and has been shown to depend on the ubiquitin-conjugating enzyme Pex4p, its membrane anchor Pex22p and the presence of the peroxisomal RING–peroxin complex (Pex2p, Pex10p, Pex12p). The ubiquitination is thought to prime the PTS-receptors for the recognition by the AAA-type ATPase complex (Pex1p, Pex6p), which functions as dislocase by extracting the modified PTS-receptors from the membrane. Furthermore, the recent “export-driven-import model” postulates that the ATP-dependent export of the PTS-receptors may be directly linked to the translocation of the cargo proteins into the peroxisomal matrix. In case the monoubiquitination-dependent receptor recycling pathway is blocked, the PTS-receptors become substrates of a polyubiquitination-dependent proteolytic pathway, which promotes their degradation by the 26S proteasome.

Fig. 15.2
figure 2

Ubiquitination and export of the PTS1-receptor at the peroxisomal membrane in Saccharomyces cerevisiae. At the peroxisomal membrane, the PTS1-receptor Pex5p is either mono- or polyubiquitinated. Both ubiquitination cascades are initiated by the ATP-dependent ubiquitin-activating enzyme Uba1p (E1). For monoubiquitination of Pex5p as part of the typical receptor cycle (left), the activated ubiquitin is transferred to the ubiquitin-conjugating enzyme Pex4p (E2) and then attached to a conserved cysteine of the receptor by assistance of the RING-ligase Pex12p (E3). The ubiquitinated receptor is ATP-dependent export of Pex5p catalyzed by the Pex15p-anchored AAA-peroxins Pex1p and Pex6p. For a new round of import the ubiquitin is cleaved off by the deubiquitinating enzyme Ubp15p. For polyubiquitination of Pex5p as part of a quality control pathway (right), the activated ubiquitin is transferred to the ubiquitin-conjugating enzyme Ubc4p (E2) and then delivered to conserved lysines of the receptor by the RING-ligase Pex2p (E3). The following export of Pex5p is also performed by the AAA peroxins Pex1p and Pex6p and the polyubiquitination signal leads to a degradation of Pex5p by the 26S proteasome

Full size image

3 The Peroxisomal Receptor Export Machinery: The Exportomer

The import of matrix proteins depends on the cycle of the PTS-receptors, which itself relies on the function of each constituent of the membrane-bound peroxins. Previous studies elucidated the composition of the peroxisomal membrane-bound subcomplexes, the docking- and the RING complex (Agne et al. 2003; Hazra et al. 2002). These two complexes were referred to as the “importomer” because both are physically connected by Pex8p in S. cerevisiae and both are required for matrix protein import (Agne et al. 2003). However, later work revealed that also constituents of the AAA complex and Pex4p complex could be co-purified with the importomer components, which strongly suggests that all membrane-associated peroxin complexes required for matrix protein import are dynamically interconnected (Oeljeklaus et al. 2012; Rosenkranz et al. 2006; Platta et al. 2009). Therefore, an alternative approach to define functionally related subcomplexes could be based on the steps of the PTS-receptor cycle at the membrane in general and on the energy dependence in particular. The current view is that the association of the PTS-receptors with the peroxisomal membrane at the site of the importomer is ATP independent (Miyata and Fujiki 2005; Miyata et al. 2009; Oliveira et al. 2003; Platta et al. 2005). However, the RING complex (Pex2p, Pex10p, Pex12p) as well as the Ubc components (Pex22p, Pex4p, Ubc4p family, UbcH5 family) belong to the ATP-dependent ubiquitination cascade (El Magraoui et al. 2012; Grou et al. 2008; Platta et al. 2007a, b, 2009; Williams et al. 2007, 2008, 2012; Liu and Subramani 2013; Kaur et al. 2013) and form together with the AAA-type ATPase complex (Pex1p, Pex6p, Pex15p, Ubp15p, AWP1) (Debelyy et al. 2011; Hensel et al. 2011; Leon et al. 2006b; Miyata and Fujiki 2005; Miyata et al. 2012) the receptor export machinery, or alternatively, peroxisomal exportomer (Table 15.1).

Table 15.1 Peroxisomal import factors
Full size table

3.1 The Ubiquitin-Conjugating Enzymes Required for Monoubiquitination of the Receptors

Ubiquitination is a posttranslational protein modification that is mediated by a three-step enzyme cascade. The ubiquitin-activating enzyme (E1) activates ubiquitin via an AMP-bound intermediate and transfers it to an ubiquitin-conjugating enzyme (E2). Finally, an ubiquitin-protein ligase (E3) binds the ubiquitin-charged E2 as well as the substrate protein, thereby enabling the transfer of the ubiquitin moiety to the target amino acid residue of the substrate (Kerscher et al. 2006; Ravid and Hochstrasser 2008). Regularly, the epsilon-amino group of a lysine within the target protein is covalently linked to ubiquitin via an isopeptide bond. Interestingly, ubiquitin can also be attached via a peptide bond to the alpha-amino group to the N-terminal amino acid, or via an oxyester bond to a threonine or serine, or even via a thioester bond to a cysteine (Wang et al. 2012).

The peroxisomal matrix protein import depends on the unusual ubiquitination of a conserved cysteine of the PTS1-receptor Pex5p (Carvalho et al. 2007; Okumoto et al. 2011; Williams et al. 2007) and of the PTS2-co-receptors Pex18p (Hensel et al. 2011) or Pex20p (Liu and Subramani 2013). The E2-enzyme that has been demonstrated to catalyze the cysteine-dependent monoubiquitination of S. cerevisiae Pex5p both in vivo and in vitro is Pex4p (Ubc10p; Platta et al. 2007a; Williams et al. 2007). Recently, the monoubiquitination of S. cerevisiae Pex18p (El Magraoui et al. 2013) and P. pastoris Pex20p (Liu and Subramani 2013) has been demonstrated to depend on the presence of Pex4p as well.

The soluble E2-enzyme Pex4p is essential for the import of both PTS1 and PTS2 proteins and therefore was the first E2-enzyme shown to be essential for the biogenesis of an organelle (Crane et al. 1994; van der Klei et al. 1998; Wiebel and Kunau 1992; Zolman et al. 2005). Pex4p is anchored to peroxisomes via the membrane protein Pex22p (Koller et al. 1999; Zolman et al. 2005). The crystal structure of S. cerevisiae Pex4p complexed to Pex22p (without its membrane domain) revealed that the Pex22p-binding site in Pex4p does not resemble a common substrate-binding motif and therefore it has been suggested that Pex22p may act as a co-activator of this E2-enzyme (Williams et al. 2012).

The molecular function of the Pex4p-catalyzed monoubiquitination of the membrane-bound Pex5p is to prime the PTS1-receptor for export (Platta et al. 2007a). While Pex4p and Pex22p are well conserved in yeasts and plants, they are absent in the genomes of mammals (Kiel et al. 2006). Instead, members of the E2D family of E2-enzymes (UbcH5a, UbcH5b and UbcH5c) fulfill the function of Pex4p in mammals (Grou et al. 2008). They catalyze the monoubiquitination of mammalian Pex5p on the conserved cysteine and therefore are required for the receptor export in vitro (Grou et al. 2008). Even though the three UbcH5 proteins carry out a central task in peroxisome biogenesis, their cellular targets are not restricted to this organelle (Brzovic and Klevit 2006; Gonen et al. 1999; Saville et al. 2004). Future work may reveal why the monoubiquitination of Pex5p has been transferred to the promiscuous UbcH5 proteins. One possible explanation could be that they control cellular events that are interconnected with peroxisome function in a concerted manner.

3.2 The RING–Peroxin Complex

The import of peroxisomal matrix proteins requires the presence of the three RING-finger proteins Pex2p, Pex10p, and Pex12p (Albertini et al. 2001; Chang et al. 1999; Eckert and Johnsson 2003; Okumoto et al. 2000; Berteaux-Lecellier et al. 1995; Krazy and Michels 2006; Peraza-Reyes et al. 2008; Sparkes et al. 2003). They have been found to form a distinct subcomplex at the peroxisomal membrane (Agne et al. 2003; Hazra et al. 2002). Defects in the assembly of the human RING complex are the second most common cause of peroxisomal biogenesis disorders (Ebberink et al. 2011; Steinberg et al. 2006). The mammalian Pex2p (formerly PAF-1) was the first gene that could be linked to PBDs (Shimozawa et al. 1992; Tsukamoto et al. 1991). Work from S. cerevisiae and A. thaliana in recent years has uncovered that all three peroxins display ubiquitin-protein ligases activity (Kaur et al. 2013; Platta et al. 2009; Williams et al. 2008). The RING–peroxins are directly involved in the ubiquitination of the PTS1-receptor Pex5p in S. cerevisiae (Platta et al. 2009; Williams et al. 2008). Their activity is also required for the formation of ubiquitinated Pex20p in P. pastoris (Liu and Subramani 2013).

In general, E3-enzymes determine the substrate specificity of ubiquitination reactions because they bind the ubiquitin-charged E2-enzyme and the substrate, thereby insuring a specific transfer of ubiquitin to the target amino acid. RING-type E3-enzymes catalyze the direct transfer of ubiquitin from the E2-enzyme to the substrate (Deshaies and Joazeiro 2009). They belong to the superfamily of Treble-Clef fold-containing proteins. This scaffold structure, which is stabilized by a Zn2+-ion, functions as an interaction motif in diverse proteins even outside the ubiquitin system (Burroughs et al. 2011). The canonical RING-finger domain (Freemont et al. 1991) binds two Zn2+-ions through its conserved Cys and His residues in a “cross-brace” arranged manner (Deshaies and Joazeiro 2009). The RING domains of Pex2p and Pex10p coordinate two Zn2+-ions, whereas the RING-finger of Pex12p binds only one Zn2+-ion (Koellensperger et al. 2007). It is interesting to note that also several members of the RBR-(RING-between-RING) family of E3-enzymes, e.g., the Parkin-like Ariadne, contain an active RING domain at their carboxy-terminus containing a single Zn2+-ion (Eisenhaber et al. 2007).

The RING–peroxins Pex2p, Pex10p, and Pex12p assemble to a distinct complex and stabilize each other in vivo (Agne et al. 2003; Hazra et al. 2002). Based on earlier binary interaction studies (Albertini et al. 2001; Chang et al. 1999; Eckert and Johnsson 2003; Okumoto et al. 2000) and recent in vitro interaction data on all three RING domains (El Magraoui et al. 2012), the RING–peroxins are thought to form a heterotrimeric complex. Pex10p (RING) functions as central component of the ternary complex as it directly binds to Pex2p(RING) and Pex12p(RING) thereby bridging the indirect interaction between these two RING domains (El Magraoui et al. 2012). The heteromeric architecture of the RING complex has a direct influence on the E3-ligase activity of the RING–peroxins because the ubiquitination activity of the combined Pex10p/Pex12p RING-domains is enhanced in presence of Pex4p in vitro (El Magraoui et al. 2012).

Pex10p also fulfills additional tasks that are distinct from Pex2p and Pex12p. A systematic functional screen of all peroxins in A. thaliana uncovered that only Pex10p has a pleiotropic growth phenotype (Nito et al. 2007). Furthermore, overexpression experiments of proteins with mutated RING domain in wild-type background suggested that A. thaliana Pex10p but not Pex2p or Pex12p are required for the contact of peroxisomes to chloroplasts during photorespiration (Prestele et al. 2010; Schumann et al. 2007). However, whether this association is due to a physical interaction or due to a functional interaction via ubiquitination events remains to be investigated.

Early studies already linked the function of the RING–peroxins to the recycling of Pex5p (Chang et al. 1999; Dodt and Gould 1996) and Pex20p (Leon et al. 2006b) as these receptors accumulate at the peroxisomal membrane in cells with disrupted RING complex. Because the monoubiquitination of Pex5p is reported to be essential for the export of Pex5p (Grou et al. 2008; Platta et al. 2007a; Okumoto et al. 2011) and as Pex12p (RING) cooperates with Pex10p(RING) in vitro (El Magraoui et al. 2012) and catalyzes this Pex4p-dependent modification in vivo (Platta et al. 2009), the Pex10p/Pex12p unit may function as the physiologic active ligase complex dedicated to the monoubiquitination-mediated export of the PTS1-receptor.

3.3 The Peroxisomal AAA-Type ATPase Complex

The ubiquitinated PTS1-receptor Pex5p is substrate for the peroxisomal AAA-type ATPase complex, which functions as dislocase that extracts Pex5p from the membrane and thereby exports it back to the cytosol (Fujiki et al. 2012; Grimm et al. 2012; Miyata and Fujiki 2005; Platta et al. 2005, 2008). The two peroxisomal AAA proteins Pex1p and Pex6p display a non-redundant and essential function in this process (Birschmann et al. 2005; Kiel et al. 1999, 2000; Tamura et al. 1998; Tamura et al. 2006).

The AAA peroxins associate with peroxisomes via an interaction of Pex6p to the tail-anchored membrane protein Pex15p in yeast and the orthologous Pex26p in mammals as well as APEM9 in plants (Birschmann et al. 2003; Furuki et al. 2006; Goto et al. 2011; Matsumoto et al. 2003a, b). An impaired assembly of the human AAA complex is the most common cause of Zellweger syndrome spectrum disorders (Geisbrecht et al. 1998; Steinberg et al. 2006). It is interesting to point out that Pex1p (formerly PAS1) was the first peroxin to be identified and also one of the founding members of the AAA family (Beyer 1997; Erdmann et al. 1991; Kunau et al. 1993).

In general, AAA proteins are characterized by a conserved modular architecture. They can be classified as P-loop NTPases, which are characterized by conserved motifs for NTP binding (Walker A motif) and hydrolysis (Walker B motif; Walker et al. 1982). AAA proteins in particular are defined by the evolutionary conserved AAA domain that contains the Walker A and B motifs as well as other conserved regions like the Second Region of Homology (SRH; Beyer 1997; Neuwald et al. 1999; Wendler et al. 2012). Pex1p and Pex6p harbor two AAA domains (AAA-D1 and AAA-D2) as well as an N-terminal domain (NTD). The binding and hydrolysis of ATP by the AAA peroxins are thought to result in conformational changes, as shown for p97 (Beuron et al. 2003), ClpX (Stinson et al. 2013) or NSF (Cipriano et al. 2013).

Most AAA proteins form active oligomers with predominantly hexameric constitution (Iyer et al. 2004). However, the current knowledge on the structural assembly of the AAA peroxins Pex1p and Pex6p is still scarce and even though they are thought to form a hetero-oligomeric complex, the stoichiometry has not yet been solved. Distinct ATP-binding and hydrolysis sites contribute to the assembly of the AAA complex (Birschmann et al. 2003, 2005; Nashiro et al. 2011; Tamura et al. 2006; Saffian et al. 2012). In yeast, ATP binding and hydrolysis in Pex6p regulate the assembly and disassembly with Pex15p (Birschmann et al. 2003), while the Pex1p–Pex6p interaction is influenced by ATP binding in D2 of Pex1p (Birschmann et al. 2005). Furthermore, the release of the AAA peroxins from the peroxisomal membrane might be regulated by the E2-enzyme Pex4p, because Pex1p and Pex6p accumulate at the peroxisome in Pex4p-deficient yeast cells. This might indicate that the ubiquitin-dependent PTS1-receptor cycle and the dynamic ATPase cycle of the AAA peroxins are interconnected (Rosenkranz et al. 2006).

In addition to their involvement in matrix protein import, the AAA peroxins have been suggested to function in the fusion of pre-peroxisomal vesicles in yeasts (Titorenko and Rachubinski 2000; van der Zand et al. 2012), while Pex6p seems to be involved in the suppression of different cell death mechanisms (Jungwirth et al. 2008; Seo et al. 2007; Warner et al. 2003). However, the best analyzed function of Pex1p and Pex6p to date is their role in peroxisomal matrix protein import (Miyata and Fujiki 2005; Platta et al. 2005; Grimm et al. 2012; Fujiki et al. 2012).

While accumulating evidence strongly indicates that the purpose of monoubiquitination is to prime Pex5p for AAA complex-mediated dislocation, the direct mechanistic purpose of this modification remains elusive. In this context, it is interesting to note that the X-ray structure of the N-domain of murine Pex1p contains a double-psi-beta-barrel fold (Shiozawa et al. 2004). This fold is also present in the N-domain of p97, where it functions as binding module for ubiquitin (Park et al. 2005). However, if the domain found in Pex1p carries out a similar function still has to be investigated. AWP1 (Associated with PRK1) has been identified as a novel binding protein of human Pex6p (Miyata et al. 2012) and is supposed to contribute to linking of the AAA peroxins to the ubiquitinated Pex5p. Accordingly, AWP1 is required for peroxisomal biogenesis in vivo and the protein interacts with both Pex6p as well as with monoubiquitinated Pex5p (Miyata et al. 2012). Thus, AWP1 might function as specific adaptor, which links the modified Pex5p to the AAA peroxins and enables them to transfer their suggested pulling force to the monoubiquitinated PTS1-receptor. Interestingly, AWP1 has also been described as an ubiquitin-binding modulator of NF-kappaB (Fenner et al. 2009).

3.4 Deubiquitination of the Receptor

The ubiquitin moiety is removed from the PTS1-receptor during or shortly after the export step but certainly prior to a new round of matrix protein import. In general, the cleavage of ubiquitin from a substrate protein is catalyzed by ubiquitin hydrolases that are called deubiquitinating enzymes (Amerik and Hochstrasser 2004). The ubiquitin hydrolase Ubp15p has been identified as a binding partner of Pex6p in S. cerevisiae (Debelyy et al. 2011). Ubp15p functions as deubiquitinating enzyme acting on Pex5p, which represents the first characterized target of this enzyme (Debelyy et al. 2011). Work based on an in vitro system with mammalian proteins suggests that the thioester bond between ubiquitin and Pex5p can be cleaved either non-enzymatically via a nucleophilic attack of glutathione or, as the major pathway, enzyme-catalyzed by ubiquitin hydrolases (Grou et al. 2009b). USP9X has been described as the main deubiquitinating enzyme acting on mammalian Pex5p (Grou et al. 2012). USP9X is a cytosolic protein whose function is not restricted to peroxisomal protein import because it has been described to take part in the regulation of the transforming growth factor beta (TGFbeta) pathway (Dupont et al. 2009).

4 Functional Link Between Receptor Export and Cargo Release

Early work has defined that the import of peroxisomal matrix proteins requires the hydrolysis of ATP (Imanaka et al. 1987). Later studies identified the export of the receptor back to the cytosol as the energy-dependent step (Oliveira et al. 2003; Gouveia et al. 2003). In recent years, it has become evident that the ubiquitination machinery (Carvalho et al. 2007; Grou et al. 2008, 2009b; Okumoto et al. 2011; Platta et al. 2007a) as well as the AAA complex (Leon et al. 2006b; Miyata and Fujiki 2005; Miyata et al. 2012; Platta et al. 2005, 2007a; Kerssen et al. 2006) can be regarded as the only ATP-consuming factors of the peroxisomal protein import machinery. This indicates that energy consumption, matrix-protein import, and PTS-receptor export merge at the exportomer.

In this respect, it is interesting to note that the protein composition of the exportomer is functionally and evolutionary related to the proteins of the endoplasmic reticulum associated degradation (ERAD) machinery (Gabaldon et al. 2006; Schluter et al. 2006). ERAD can be defined as a mechanism by which misfolded proteins are polyubiquitinated and extracted from the ER in order to be disposed by the 26S proteasome in the cytosol (Hampton and Sommer 2012). Translocation systems that are in many aspects comparable to the exportomer and ERAD are the mitochondria associated degradation (MAD) for proteins of the outer mitochondrial membrane (Taylor and Rutter 2011) as well as the pre-protein translocator of complex plastids called symbiont-derived ERAD-like machinery (SELMA; Bolte et al. 2011). Therefore, a mechanistic parallel can be drawn between the exportomer, ERAD-, MAD-, and SELMA substrates because all are extracted by mechanoenzymes of the AAA-type ATPase family in an ubiquitination-dependent manner (Bolte et al. 2011; Platta et al. 2007b; Schliebs et al. 2010).

Based on this similarity, a model has been proposed that draws a direct interconnection of receptor export and the translocation of matrix proteins across the peroxisomal membrane (Schliebs et al. 2010). This “export-driven import model” is supported by the fact that the presence of a functional exportomer is a prerequisite for the import of matrix proteins. This, ATP is required for the ubiquitin- and AAA-driven extraction of the receptor and might be mechanically coupled to the translocation of the cargo proteins over the membrane.

Accordingly, the import defects observed in mutants of the exportomer can be explained in two ways. First, the binding capacity for functional PTS-receptors at the peroxisomal membrane seems to be limited. In fact, a decreased rate of receptor export caused by the functional impairment of the export machinery leads to an accumulation of PTS-receptors at the membrane (Leon et al. 2006b; Platta et al. 2004) and therefore would block the docking of new receptor–cargo complexes from the cytosol. In A. thaliana, the physiological defects of mutated and only insufficiently active Pex6p could be partially overcome when it was co-expressed with a weak allele of the docking protein Pex13p (Ratzel et al. 2011). This finding strongly indicates that the import and export rates of the PTS-receptors need to be balanced. Second, this model suggests that export of the receptor and the release of the cargo might be directly linked by a concerted mechanism. Work on the ubiquitination of the S. cerevisiae PTS2-co-receptor Pex18p delivered first direct evidence for such a connection (Hensel et al. 2011). Based on protease-protection assays, it was revealed that Pex7p is partially protease protected in wild-type cells, while Pex18p remains accessible. This topology is reversed when the cysteine of Pex18p is mutated or the AAA peroxins Pex1p/Pex6p are deleted (Hensel et al. 2011). This finding strongly indicates that monoubiquitination of Pex18p as well as AAA complex governs the import of cargo-loaded Pex7p. However, in the mammalian system, it is not yet clear whether the cargo release step itself requires ATP hydrolysis (Miyata et al. 2009) or does not (Alencastre et al. 2009).

In conclusion, the receptor export machinery is thought to function as the energy-consuming import motor for matrix proteins, either indirectly via balanced receptor import/export rates and/or directly via an interconnection of receptor export and cargo translocation.

5 Polyubiquitination of the PTS-Receptors

Under certain conditions, the PTS-receptors Pex5p, Pex18p, and Pex20p are polyubiquitinated on lysine residues in order to mark them for the degradation by the 26S proteasome (Hensel et al. 2011; Kiel et al. 2005b; Leon et al. 2006b; Platta et al. 2007a; Williams et al. 2007). This proteolytic pathway is induced when the normal monoubiquitination-dependent recycling pathway is blocked, as it is the case when constituents of the Pex4p or AAA complexes are deleted or the conserved cysteine of the PTS-receptor is mutated (Kiel et al. 2005a; Kragt et al. 2005; Platta et al. 2004; Leon and Subramani 2007; Hensel et al. 2011). The polyubiquitination of the S. cerevisiae PTS1-receptor Pex5p is predominantly catalyzed by Ubc4p and to a minor portion by the partial redundant enzymes Ubc5p and Ubc1p (Kiel et al. 2005a; Kragt et al. 2005; Platta et al. 2004). These three ubiquitin-conjugating enzymes display a high sequence similarity and are involved in diverse other cellular processes as well (Seufert and Jentsch 1990; Seufert et al. 1990). Both Pex10p (Williams et al. 2008) as well as Pex2p (Platta et al. 2009) have been suggested to function as E3 enzymes for the polyubiquitination of Pex5p. In this respect, it is interesting to note that a recent in vitro study demonstrates that Pex10p (RING) can synergistically enhance the ubiquitination activity of the Ubc4p–Pex2p (RING) enzyme pair (El Magraoui et al. 2012). This result suggests that both RING–peroxins may act together in the Ubc4p-dependent generation of K48-linked polyubiquitin chains on Pex5p.

Receptor polyubiquitination is enhanced when the export machinery is affected in its function and therefore the purpose of this modification is likely to remove the aberrant receptor molecules from the membrane when the normal extraction and recycling reaction is hampered. However, mutagenesis of the lysine residues required for polyubiquitination of S. cerevisiae Pex5p does not lead to a growth defect on oleate medium (Platta et al. 2007a; Williams et al. 2007). Interestingly, polyubiquitination of Pex5p can also be regarded as an alternative export signal. In vitro export assays demonstrated that a fraction of Pex5p is still exported even in a Pex4p-deficient system when the two conserved lysine residues required for polyubiquitination were still present (Platta et al. 2007a). Moreover, mutation of the conserved cysteine in P. pastoris Pex20p (Leon and Subramani 2007) induces polyubiquitination of Pex20p but still retains a partial functional receptor molecule that displays partial complementation in growth tests. Interestingly, only both the non-essential lysine targets for polyubiquitination of Pex20p as well as to the typically monoubiquitinated cysteine are mutated, the receptor completely loses its functionality (Leon and Subramani 2007). These data demonstrate that the enhanced degradation of Pex20p can restore the matrix protein import to a certain extent, supposedly because the receptors are removed efficiently enough to allow the docking of further cargo-bound receptors. This mechanism has been described as RADAR (receptor accumulation and degradation in the absence of recycling; Leon et al. 2006a, b) in order to distinguish it from the non-essential quality control. However, it should be noted that the mutation of the conserved cysteine of S. cerevisiae Pex5p and Pex18p alone is already sufficient to fully abolish the function of these receptors (Hensel et al. 2011; Williams et al. 2007). In this respect, it is interesting to note that degradation of Pex5p occurs much slower in S. cerevisiae than in most other species (Collins et al. 2000; Dodt and Gould 1996; van der Klei et al. 1998; Zolman and Bartel 2004; Zolman et al. 2005). Therefore, the observed instability of Pex5p in exportomer mutants in these species is most likely due to rapid degradation via K48-linked polyubiquitination as described for the PTS1-receptor of H. polymorpha (Kiel et al. 2005b).

The S. cerevisiae PTS2-co-receptor Pex18p behaves somewhat different from the PTS2-co-receptor Pex20p in P. pastoris and H. polymorpha because Pex18p shows a constitutive turnover already under wild-type conditions (Hensel et al. 2011; Leon et al. 2006b; Otzen et al. 2005; Purdue and Lazarow 2001). Currently, the functional impact of this instability is not known. In contrast to Pex18p, the PTS2-receptor Pex7p of S. cerevisiae is a stable protein (Hensel et al. 2011). So far, no indications for an ubiquitination of yeast Pex7p have been found. Interestingly, a recent report describes the polyubiquitination and degradation of Arabidopsis Pex7p when the dominant-negative GFP-Pex7p species is expressed in the cell (Cui et al. 2013). However, it is not clear if this mechanism is conserved in other organisms.

In general, the removal of the PTS-receptors via polyubiquitination is initiated when the monoubiquitination-dependent recycling pathway is blocked and therefore may function as alternative export signal.

6 Concluding Remarks

The combined work of several laboratories on the ubiquitination and recycling of the PTS-receptors has helped to uncover the functional contribution of distinct peroxisomal subcomplexes to the dislocation step and therefore enabled the definition of the peroxisomal receptor export machinery, the exportomer (Platta et al. 2013).

Certainly, many open questions remain to be answered and one of the most intriguing ones concerns the finding that Pex5p, Pex18p, and Pex20p are monoubiquitinated on a cysteine via a thioester bond and not by a more common isopeptide bond to a lysine. The first evidence that ubiquitin can be attached to cysteine, serine or threonine residues came from studies of viral MARCH (Membrane-associated RING-CH) E3 ligases that ubiquitinate MHC I (Major Histocompatibility Complex I) molecules (Cadwell and Coscoy 2005; Wang et al. 2007), and recent work demonstrates that this uncommon ubiquitination can also take place during ERAD (Ishikura et al. 2010; Shimizu et al. 2010). However, it is unclear how the specificity for these non-lysine ubiquitination reactions is ensured because the E2 and E3 enzymes involved are not restricted to this kind of modification and can also modify lysine residues (Wang et al. 2012). Interestingly, the cysteine of mammalian Pex5p (Grou et al. 2009b, 2012) and P. pastoris Pex20p (Leon and Subramani 2007) can be replaced by a lysine, which results in a still largely functional protein. Thus, even though the cysteine and the thioester-bond mediated ubiquitination of the PTS-receptors are evolutionary conserved, they are not essential for the principle export mechanisms and therefore may mainly represent an important regulatory device.

There are several different possibilities to explain the function of the conserved cysteine of the peroxisomal receptors. (1) The first concept is based on the fact that thioester bonds are less stable in comparison to isopeptide bonds. Therefore, the duration of the ubiquitin moiety at the PTS-receptor might be restricted in order to disable the formation of a polyubiquitin chain or to prevent the recognition by proteasomal adaptors. The rapid non-enzymatic disruption of the thioester bond of Ub-Pex5p in a mammalian in vitro system supports the idea that the cysteine-ubiquitination protects the PTS-receptors against degradation (Grou et al. 2009a, b). (2) Another concept takes into account that that certain E3 enzymes, like HECT-type ligases (Kee and Huibregtse 2007) or the RBR-type ligases (Wenzel et al. 2011) form an ubiquitin-thioester intermediate on a cysteine before this ubiquitin molecule is finally transferred to the substrate protein. One hypothetical model could be that once Pex12p/Pex10p have modified one of the receptor molecules of the oligomeric pore, Pex5p itself could catalyze an intra-oligomeric ubiquitin transfer in a relay-like system in order to accelerate the decomposition of the pore (Erdmann and Schliebs 2005; Platta et al. 2013). Interestingly, the E2-enzyme E2-230 K represents an example of an intramolecular ubiquitin transfer (Berleth and Pickart 1996), where ubiquitin is transferred from the first cysteine to a second cysteine of E2-230 K prior to attachment of the ubiquitin to the target protein. (3) A third concept is related to recent work on the regulation of the peroxisomal redox balance (Ivashchenko et al. 2011), which contributes to the general functional role of peroxisomes in the control of the cellular levels of reactive oxygen species (Bonekamp et al. 2009). One possibility is that the cysteine required for monoubiquitination might be accessible for redox changes. This might have a direct impact on the availability of this residue for the monoubiquitination and therefore could control the import/export rates of the receptor (Fransen et al. 2012).

In conclusion, the understanding of the late acting peroxins as concerted acting components of the exportomer will be instrumental to uncover the molecular mechanism underlying peroxisomal matrix protein import.