SEARCH

SEARCH BY CITATION

Keywords:

  • copper chaperone of SOD1;
  • peroxisomes;
  • piggyback;
  • protein import

Abstract

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Most newly synthesized peroxisomal proteins are imported in a receptor-mediated fashion, depending on the interaction of a peroxisomal targeting signal (PTS) with its cognate targeting receptor Pex5 or Pex7 located in the cytoplasm. Apart from this classic mechanism, heterologous protein complexes that have been proposed more than a decade ago are also to be imported into peroxisomes. However, it remains still unclear if this so-called piggyback import is of physiological relevance in mammals. Here, we show that Cu/Zn superoxide dismutase 1 (SOD1), an enzyme without an endogenous PTS, is targeted to peroxisomes using its physiological interaction partner ‘copper chaperone of SOD1’ (CCS) as a shuttle. Both proteins have been identified as peroxisomal constituents by 2D-liquid chromatography mass spectrometry of isolated rat liver peroxisomes. Yet, while a major fraction of CCS was imported into peroxisomes in a PTS1-dependent fashion in CHO cells, overexpressed SOD1 remained in the cytoplasm. However, increasing the concentrations of both CCS and SOD1 led to an enrichment of SOD1 in peroxisomes. In contrast, CCS-mediated SOD1 import into peroxisomes was abolished by deletion of the SOD domain of CCS, which is required for heterodimer formation. SOD1/CCS co-import is the first demonstration of a physiologically relevant piggyback import into mammalian peroxisomes.

Peroxisomes are organelles involved in numerous anabolic and catabolic pathways (1). They contain various oxidases that generate H2O2 in addition to a number of enzymes converting reactive oxygen species into non-toxic metabolites (2). Recently, it was shown that peroxisomes also produce superoxide radicals (inline image) (3). As a superoxide scavenger, Cu/Zn superoxide dismutase 1 (SOD1), a protein mutated in the fatal disease familial amyotrophic lateral sclerosis (4), metabolizes inline image into O2 and H2O2, the substrate of the major peroxisomal enzyme catalase. Several groups (5–10) have proposed that SOD1 is a peroxisomal protein; however, the enzyme could also be found in the cytosol or in mitochondria (11). The mechanisms for targeting SOD1 to distinct subcellular compartments are unknown. Most peroxisomal proteins are imported into the organellar matrix via either the Pex5 or Pex7 import receptor-mediated pathways depending on different targeting sequences at the C-terminal [peroxisomal targeting signal 1 (PTS1)] or the N-terminal (PTS2) ends of the polypeptide chain, respectively (12). However, SOD1 and several other peroxisomal proteins like lactate dehydrogenase (LDH) (13) or the liver fatty acid binding protein (14) harbor neither a PTS1 nor a PTS2, and it is still enigmatic how these proteins enter the peroxisomal matrix. Based on the experimental data using artificial reporter molecules, an import in a piggyback fashion was proposed for peroxisomal proteins without targeting signals using proteins carrying a PTS as a shuttle (15). Although this piggyback hypothesis was proposed for more than a decade, a physiological example in higher organisms could not be identified so far. Therefore, the relevance of such a pathway has recently been questioned (16). The detection of both SOD1 as well as copper chaperone of SOD1 (CCS) in isolated peroxisomes using mass spectrometry (MS) (17) led us to the assumption that SOD1 could be imported into peroxisomes using its interaction partner CCS (18) as a shuttle, which bears a putative PTS1. In this study, we show by immunoblotting of isolated subcellular fractions as well as immunocytochemistry and immunoelectron microscopy that SOD1 and CCS are indeed true peroxisomal residents. Furthermore, SOD1 that is predominantly localized in the cytosol could be enriched in peroxisomes by the overexpression of CCS in CHO cells. This effect was abolished after deletion of the PTS1 of CCS or its interaction domain to SOD1, respectively. These results provide the first evidence for a naturally occurring piggyback import mechanism in mammalian cells.

Results

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

In the course of our recent analysis of the peroxisomal proteome, we could confirm that SOD1 is indeed present in highly purified rat liver peroxisomes (purity >95%) (17). In addition, its copper chaperone CCS, which directly interacts with SOD1 to exchange Cu2+ ions (18), was also detected in the same peroxisomal fraction. Both proteins were identified by three peptides using an Applied Biosystems 4700 Proteomics Discovery system. A potential overlap of the identified peptides because of a ca. 50% homology of their primary sequence was excluded by manual inspection of the peptide sequences. As shown in Table 1, SOD1 and CCS were identified by peptides with significant mass differences, resulting in individual amino acid sequences characteristic for each protein. Because Casareno et al. (18) did not find a peroxisomal localization of CCS in HepG2 cells using immunocytochemical staining, we reinvestigated the intracellular distribution of both proteins in rat liver using immunoblotting of subcellular fractions as well as immunoelectron microscopy of rat liver sections. For this purpose, a polyclonal antiserum against rat CCS was raised in guinea pigs. The antibody fraction purified from the serum showed a strong reaction to a single protein band of approximately 33 kDa in cytosol as well as peroxisomes [calculated molecular weight (MW): 29 kDa]. This band disappeared, when the antibody was preincubated with the peptide used for immunization (Figure 1A). To compare the localization of CCS and SOD1, different subcellular fractions were isolated from liver tissue, and western blots for the immunodetection of SOD1 and CCS as well as of characteristic organellar marker proteins were performed (Figure 1B). The mitochondrial and cytosolic fractions used showed organellar cross contaminations as shown in Table 2. Endoplasmic reticulum (ER), lysosomal and nucleic fractions were only used as internal negative controls and were not completely enzymatically characterized. SOD1 was localized to several compartments (Figure 1B). The highest concentration was detected in the cytosol; but, in addition, SOD1 was also found in mitochondria, where it was described as a resident of the intermembrane space (11), and in peroxisomes. In this respect SOD1 resembles LDH, another protein with an established distribution to several compartments (13). In contrast to SOD1, CCS was detectable in approximately equal concentrations in the cytosol and in peroxisomes, but to a much lesser extent in mitochondria. To further corroborate the localization in peroxisomes, we performed immunoelectron microscopy in slices of rat liver tissue. Immunolabeling showed that the peroxisomal marker enzyme catalase was abundant and evenly distributed throughout the peroxisomal matrix (Figure 2A). Beside its well-described cytoplasmic localization, CCS was also specifically detected within the peroxisomal matrix but with a lower abundance than catalase (Figure 2B). In addition, mitochondria were sporadically labeled (Figure S1A). SOD1 showed a similar peroxisomal staining as CCS (Figure 2C), but was found with a significantly higher labeling density in the cytoplasm (Figure S1B). Control incubations without primary antibodies against catalase, CCS or SOD1 did not result in any specific labeling (not shown). To verify the specificity of the peroxisomal CCS labeling, we also used peptide-depleted antiserum for immunoelectron microscopy. The labeling density was significantly decreased (p < 0.001) after incubation with depleted antiserum (0.37 gold particles per peroxisome) compared with non-depleted antiserum (1.45 gold particles per peroxisome). Furthermore, CCS labeling density was compared in different subcellular compartments. Our analysis revealed highest labeling densities for CCS in peroxisomes and in the cytosol with 6.86 and 3.62 gold particles per μm2, respectively, further corroborating the data obtained by western blot analysis of highly purified organelle fractions (Figure 1B). In contrast, the labeling densities for mitochondria, rough endoplasmic reticulum (rER) and the nucleus were only 1.92, 1.87 and 1.85 gold particles per μm2, respectively. The detection of SOD1 and CCS in peroxisomes by MS as well as immunological methods provides strong evidence that both proteins are true peroxisomal residents. The fact that CCS harbors a C-terminal PTS1 (AHL), according to the PTS1 predictor analysis software (19), raised the question whether the peroxisomal localization of both proteins and their known interaction might provide the basis for a co-import of both proteins as a complex. In that case, CCS would be responsible for the peroxisomal import of SOD1 in a piggyback fashion.

Table 1.  Mass spectrometrical identification of SOD1 and CCS in peroxisomes.
 Calculated massObserved massSequenceIon scoreCI (%)
  1. CI, confidence interval.

SOD1  P0763280100.0
 1383.72271383.6958KHGGPADEER2897.51
 1425.80591425.6699HVGDLGNVAAGK2796.78
 1798.96971798.7974GDGPVQGVIHFEQG2595.81
CCS  Q9JK7275100.0
 1055.61941055.5643GRPIAGQGR3399.19
 1095.60781095.5076GMGSSQLK2796.81
 1426.71581426.6000GDLGNVHAEASGR1663.67
image

Figure 1. Detection of SOD1 and CCS in different subcellular fractions. A) Validation of the guinea pig antibody against CCS. With the preimmune serum, no specific bands except for a faint staining of the abundant peroxisomal proteins catalase and uricase could be detected. In contrast, the serum after immunization stains a single band in both cytosol and peroxisomes of approximately 33 kDa, which disappears after a preincubation with the peptide used for immunization. B) Cellular distribution of CCS and SOD1. Using immunoblotting, CCS can be localized to the cytosol as well as to peroxisomes isolated from rat liver. SOD1 is most abundant in the cytosol, but can also be found to minor extents in mitochondria and peroxisomes. Lactate dehydrogenase (LDH), another protein with a partial peroxisomal localization, shows a similar distribution. Catalase (peroxisomes), ATP synthase (mitochondria), ERp29 (microsomes) and lamin 2 (nucleus) were used as markers for the corresponding cell fractions. Five micrograms of total protein was applied per lane.

Download figure to PowerPoint

Table 2.  Properties of mitochondrial and cytosolic fractions.
EnzymeMitochondrial fractionCytosolic fraction
Relative specific activityPurity/ contamination (%)Relative specific activityPurity/ contamination (%)
  1. aValues correspond to the mean of two experiments; all other activities were determined at least 3×.

  2. bmeasured as acid phosphatase activity

  3. cmeasured as β-glucuronidase activity

  4. Values given are means ± SD, relative specific activity (RSA) = units/mg Po protein × (units/mg liver protein)−1 and purity/ contamination = p× RSA, where p is the percentage of total liver protein contributed by peroxisomes (2.53), mitochondria (20.2), endoplasmic reticulum (21.5) and lysosomes (2.03).

Cytochrome c oxidase3.79 ± 0.3576.560.002 ± 0.0040.04
Lactate dehydrogenase0.06a 2.16a 
Catalase0.22 ± 0.020.561.02 ± 0.152.58
Lipid β-oxidationn.d.n.d.0.39 ± 0.050.99
Acid phosphataseb/β-glucuronidasec0.62 ± 0.1b1.230.21 ± 0.07c0.42
Esterase0.35 ± 0.067.50.15 ± 0.050.15
image

Figure 2. Detection of SOD1 and CCS by immunoelectron microscopy. Ultrastructural localization of proteins was performed with specific antibodies to catalase (A), CCS (B) and SOD1 (C) on sections of rat liver. Immune complexes were visualized by gold particles [12 nm in (A) and (C) and 10 nm in (B), see arrowheads]. All three proteins were localized within the peroxisomal matrix. The crystalloid core (indicated by *) provides ultrastructural evidence for the peroxisomal nature of the organelles shown. The bar in (C) represents 500 nm.

Download figure to PowerPoint

To prove this hypothesis, we first compared the distribution of CCS and SOD1 between cytosol and peroxisomes. A high molar excess of SOD1 to CCS was consistently found in cellular homogenates of different organisms and tissues (20,21). Considering a 1:1 interaction of CCS and SOD1 (22), a piggyback transport model would only be reasonable if large amounts of SOD1 remain in the cytosol and only a smaller fraction, comparable to the quantity of peroxisomal CCS, would be imported into the organelles. In a quantitative approach, dilution series of peroxisomal and cytosolic protein fractions were analyzed for SOD1 and CCS by immunoblotting. Intensities of the corresponding signals were determined by densitometry, and the ratios of SOD1 and CCS in the cytoplasmic to peroxisomal fractions (Figure 3A) were calculated using these titration curves. For reference, ratios were also determined for catalase, a peroxisomal marker enzyme known to easily leak out of peroxisomes during the isolation procedure. A peroxisomal to cytosolic concentration ratio of about 25:1 was calculated, indicating that catalase was still highly concentrated in the isolated peroxisomes irrespective of its known leakiness. Taking into account the distribution of total cellular protein in peroxisome and cytosol fractions (2.5 versus 40%) (23), this implicates that about 35% of total catalase was released during the isolation procedure, which is in line with former publications (24). The dilution series revealed approximately equal concentrations of CCS in peroxisomal and cytosolic fractions. Referring to the release of about 35% of peroxisomal soluble matrix protein into the cytoplasm during the isolation procedure, about 8.5% of the total cellular CCS was obviously translocated to peroxisomes, while 91.5% of the protein remained in the cytosol. This limited import of CCS into peroxisomes suggests that the C-terminal sequence of CCS has to be considered as a weak PTS1 in rat.

image

Figure 3. Distribution and native conformation of SOD1 and CCS in peroxisomes and cytosol. A) Determination of the ratio between the cytosolic and peroxisomal populations of SOD1 and CCS. Dilution series of isolated peroxisomal and cytosolic fractions were analyzed using immunoblotting. CCS showed an approximately equal distribution between both subcellular compartments, whereas for SOD1 a ratio of 1:60 between peroxisomes and cytosol was calculated using densitometry of the western signals. For comparison, signals of catalase are shown to exclude peroxisomal leakage. B–E) Immunodetection of SOD1, CCS, catalase and Pex5 after Blue Native (BN)-PAGE of cytosolic or peroxisomal protein fractions (the Coomassie stained membrane—blue spots—is underlayed for orientation). In both cellular compartments, CCS migrates with the tail of monomeric proteins (B, C). SOD1, however, can be found in a protein complex not separated in the first native dimension. Note that the major SOD1 complex found in the cytosol (B) is larger than that in peroxisomes (C, arrows). In the cytosolic fraction, catalase, the most prominent peroxisomal protein in liver, was found as a complex of high molecular weight, which is in line with its described tetrameric enzymatically active state (D). However, an association of both catalase and CCS to Pex5 could not be monitored by BN-PAGE (D, E). Marker lanes shown were always run on the same gels.

Download figure to PowerPoint

In contrast to CCS, SOD1 showed a 60-fold higher concentration in the cytosol than in peroxisomes (Figure 3A) meaning that approximately 99.85% of SOD1 are distributed to the cytosol and only 0.15% to peroxisomes. Thus, while CCS was found in a comparable concentration in both compartments, SOD1 exhibits a strong bias toward a cytosolic localization. This implies that the distribution between peroxisomes and cytosol differs considerably for both proteins. Considering an excess expression of SOD1 over CCS, the diverse distribution pattern of CCS, when compared to SOD1 found in this study, supports a CCS-shuttle system, which targets only a minor fraction of cellular SOD1 to peroxisomes.

A molar excess of SOD1 in the cytosol precludes a permanent association of CCS and SOD1 to provide all SOD1 molecules with Cu2+ ions. In peroxisomes, however, a piggyback import would remodel this ratio, which could lead to a more permanent association of both proteins. To assess the stability of the SOD1/CCS complex, cytosol and peroxisome fractions of rat liver were subjected to Blue Native (BN)-PAGE. According to Figure 3B,C, the majority of SOD1 did not form a stable complex with CCS in both compartments. SOD1 was rather present as a homomeric complex or coupled to other yet unknown proteins, whereas only monomeric CCS could be detected. Furthermore, the cytosolic and peroxisomal SOD1-containing protein complexes differed slightly in their apparent MW (Figure 3B,C, arrows). To validate if an interaction of low stability would be sufficient for an efficient import of a protein into peroxisomes, we investigated the integrity of the well-described cytosolic Pex5/catalase import complex in the same BN-PAGE system. As depicted in Figure 3D, Pex5 was only found as monomer in the cytosolic fraction, whereas catalase could be detected as a stable homomeric complex, yet not in a complex with the peroxisomal import receptor Pex5. A complex between Pex5 and CCS was also not detected (Figure 5E). Thus, transient or less-stable interactions seem to be sufficient to deliver proteins to their organellar targets, whereas functional enzyme complexes as the catalase tetramer reflect more stable intermolecular associations.

image

Figure 5. The SOD domain of CCS is required for CCS-mediated SOD1 import. BGL69 cells were transfected with plasmids for expression of N-myc-CCS-D3 (A), N-myc-CCS-D1 (B) and N-myc-CCSΔPTS1 (C). The myc epitope (A–C) is labeled in red. Mixtures of plasmids for expression of N-myc-CCS-D3 (D–F), N-myc-CCS-D1 (G–I) or N-myc-CCS-D2 (J–L) together with N-flag-SOD1 were transfected into BGL69 cells and the flag epitope was immunostained in red. Peroxisomal GFP in (A–D), (G) and (J) is shown in green. Overlays of GFP fluorescence with epitope stainings are shown in (A–C), (F), (I) and (L). A localization of the epitope tags in peroxisomes results in a yellow punctate staining pattern (A, B and F). The bar in (L) represents 20 μm. A schematic illustration of expression constructs used in this study is shown in (M). The position of the epitope tags (myc and flag), the heavy metal-associated (HMA) domain, the superoxide dismutase (SOD) domain and the peroxisomal targeting signal (PTS1) are indicated. Deletions are shown as thin black lines.

Download figure to PowerPoint

Taken together, the biochemical data provided are in agreement with a CCS-mediated import of SOD1 into rat liver peroxisomes.

To better understand the import mechanism, we next analyzed the subcellular localization of epitope-tagged versions of SOD1 and of CCS after transient transfection of BGL69 cells, a CHO-derived cell line with fluorescently marked peroxisomes of low morphologic heterogeneity. CCS with an N-terminal myc-tag was almost exclusively localized to peroxisomes (Figure 4A–C). This import rate was higher than expected from our analysis of rat liver fractions, which might be because of the fact that we used modified rat proteins in our Chinese hamster ovary cell system. However, the import of myc-tagged CCS was abolished after deletion of the C-terminal AHL motif, indicating that CCS was indeed imported into peroxisomes in a PTS1-dependent fashion (Figure 5C). In contrast, a C-terminally flag-tagged SOD1 version was located in the cytoplasm and in the nucleus in the absence of CCS (Figure 4D–F) without any detectable import into peroxisomes. The same result was obtained with a SOD1 version bearing an N-terminal flag epitope (not shown). Upon coexpression of the C-terminally flag-tagged SOD1 with the N-myc-tagged CCS, however, SOD1 was detected in the cytoplasm and the nucleus and, surprisingly, in significant amounts in peroxisomes (Figure 4G–I). The same result was observed after coexpressing N-terminally flag-tagged SOD1 with myc-tagged CCS (not shown). These data indicate that SOD1 can be translocated in a substantial fraction into peroxisomes depending on the presence of CCS, but irrespective of the particular amino acid sequence of its N- or C-terminus.

image

Figure 4. Import of SOD1 into peroxisomes is mediated by CCS. BGL69 cells with GFP-labeled peroxisomes (green signal in A, D and G) were transfected with plasmids for expression of N-myc-CCS (A–C), C-flag-SOD1 (D–F) or with a mixture of both plasmids (G–I). The myc epitope (B) and the flag epitope (E and H) are shown in red. The images were obtained by confocal laser scanning microscopy. Overlays of corresponding images are shown in (C, F and I). The yellow signal indicates a colocalization of the appropriate epitope with peroxisomal GFP-PTS1. The bar in (I) represents 20 μm.

Download figure to PowerPoint

To investigate if the intracellular distribution of SOD1 depends on a direct interaction with CCS, we constructed several deletion mutants of N-myc-CCS (Figure 5M). Mutant N-myc-CCS-D1 was obtained deleting aa 137–219 of CCS, which comprise large parts of the SOD-like domain II. This leads to the loss of an intramolecular cysteine bridge (25), thereby disturbing the secondary structure of CCS. The second mutant N-myc-CCS-D2 lacked aa 183–219 of CCS. This deletion mutant preserves most of the CCS structure but disrupts one of the major contact sites between CCS and SOD1 within domain II. In addition, we tested N-myc-CCS-D3, a splice variant of CCS isolated from a rat liver cDNA preparation (GeneBank Acc. No. EU340033). This variant lacks exon 3 of CCS, leading to a deletion of aa 38–83 within the N-terminal heavy metal-associated (HMA) domain that facilitates copper binding and transfer to SOD1 (26). Upon transfection of the corresponding expression plasmids into BGL69 cells, all deletion mutants of CCS localized primarily to peroxisomes, indicating that the AHL-mediated import was not disturbed (Figure 5A,B, N-myc-CCS-D2 not shown). In contrast, a truncated CCS mutant without the C-terminal targeting information (N-myc-CCSΔPTS1, Figure 5M) was not imported into peroxisomes (Figure 5C). Upon cotransfection of the C-flag-SOD1 expression plasmid with these CCS mutants, we observed differences in the distribution pattern of SOD1. The deletion of the HMA domain of CCS in N-myc-CCS-D3, which is not essential for the interaction with SOD1 (18), did not influence the CCS-dependent peroxisomal targeting of SOD1 (Figure 5D–F). However, coexpression with both N-myc-CCS-D1 as well as N-myc-CCS-D2 resulted in an exclusively cytoplasmic localization of SOD1 without any detectable enrichment in peroxisomes (Figure 5G–L). Furthermore, C-flag-SOD1 was restrained to the cytoplasm upon coexpression with an N-myc-CCSΔPTS1 (not shown). Taken together, these findings suggest a piggyback mechanism for the translocation of SOD1 to peroxisomes for the following reasons: (i) the SOD domain of CCS, which is required for heterodimerization of SOD1 and CCS in order to exchange copper ions, is also required for the import of SOD1 into peroxisomes; (ii) the peroxisomal import of CCS and of SOD both depend on an intact targeting signal (PTS1) in the CCS amino acid sequence. In fact, our observation that both proteins were localized in the cytoplasm after cotransfection of expression plasmids in Pex5-deficient cells (Figure S2) further corroborates the importance of PTS1-mediated CCS import for co-import of SOD1 into peroxisomes.

Discussion

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

The import of homo-oligomeric complexes into peroxisomes is a well-known phenomenon, since the import of thiolase and chloramphenicol acetyltransferase in dimers and trimers into peroxisomes of Saccharomyces cerevisae or Yarrowia lipolytica has been reported (27–29). A similar mechanism has also been shown for the import of isocitrate lyase into glyoxysomes (30). In conjunction with the report of Walton et al. (31) that even gold particles conjugated to peptides bearing a PTS1 were translocated into the peroxisomal matrix, these findings led to the hypothesis of a pore-forming protein import machinery in the membrane of peroxisomes. The idea of an import of hetero-oligomeric protein complexes gained support by observations on the viral protein HIV-1 Nef (32) and a PTS1-deficient mutant of Eci1p (33), which were both shown to be translocated provided they were coupled to interaction partners bearing a PTS. It should be realized, however, that these examples reflect pathological or artificial conditions. Recently, the import of a heteropentameric acyl–CoA oxidase complex into peroxisomes of the yeast Y. lipolytica was demonstrated (34), being the first example of a physiologically relevant piggyback mechanism. In this study, we show that a similar mechanism does also exist in higher organisms showing that SOD1, a physiological cell constituent, was only directed to peroxisomes, when expressed together with CCS. Both proteins colocalized in part in peroxisomes, while another part was retained in the cytoplasm. The amount of peroxisomal SOD1 was increased by the overexpression of both proteins, indicating that the interaction between SOD1 and CCS can influence the intracellular distribution of both proteins. Interestingly, a positive impact of CCS regarding the presence of SOD1 was also reported from the intermembrane space of yeast mitochondria (35). Under these experimental conditions, however, CCS was proposed to mediate the retention of SOD1 in the mitochondrial intermembrane space by catalyzing the formation of an intramolecular cysteine bridge within SOD1, but was not directly required for the protein import into the organelle. Besides this mechanistic view of protein import, it remains tempting to speculate about the biological role of both proteins in peroxisomes. With the intraorganellar location of xanthine oxidase (36), rat peroxisomes house a long-known producer of superoxide radicals (37). In addition, Lopez-Huertas et al. (38) reported the generation of superoxide by cytochrome b5, another enzyme, which was identified as a peroxisomal constituent in mice, recently (39). Further on, because H2O2 is one of the products of the disproportionation reaction of inline image catalyzed by SOD1, a close spatial proximity of SOD1 to the peroxisomal enzyme catalase would be another physiological advantage. Regarding CCS, it is only possible to speculate on alternate functions as copper transfer to SOD1 in peroxisomes. According to its known function as a copper donator, CCS may be involved in the copper loading of other peroxisomal enzymes. At least urate oxidase was reported to contain Cu2+ as a prosthetic group (40); however, the very scarce data on the metal dependence of peroxisomal proteins comprise the potential of further Cu2+-dependent enzymes.

To our knowledge, this is the first example for the import of a physiological heterodimer into mammalian peroxisomes. Based on the extended shuttle model for import of peroxisomal proteins (41), a heterodimeric complex of SOD1 with CCS would bind to Pex5p to generate a heterotrimeric complex that is subsequently imported into peroxisomes. Reaching the peroxisomal matrix, however, both proteins dissociate to shape individual protein interactions. Generally, proteins are directed to their subcellular compartment by encoded targeting sequences (42). Our data extend this concept supporting the idea that proteins with physiological functions in several subcellular compartments could be directed to their respective location in a chaperone-mediated fashion, also in higher organisms. In addition to previous studies, we could show that a transient interaction between two proteins could be used to shuttle a PTS-deficient binding partner to peroxisomes. The differential expression of interacting partners, e.g. according to specific physiological demands, could be used for the cell-autonomous regulation of their preferential localization. In this respect, this work can give new impetus to the search for hitherto unknown piggyback relationships between PTS-bearing peroxisomal proteins and multilocalized interaction partners previously regarded as mere contaminants. This should extend the spectrum of peroxisomal functions.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Isolation of peroxisomes and other organelles

Peroxisomes were prepared as described previously (17). In brief, female Sprague Dawley rats weighing 200–250 g were anaesthetized and the livers were rapidly dissected. After removing the connective tissue, the livers were minced in an ice-cold homogenization buffer [HB, 250 mm sucrose, 5 mm 3-(N-morpholino)propanesulfonic acid (MOPS), 1 mm ethylenediaminetetraacetic acid (EDTA), 0.1% ethanol, 2 mm phenylmethylsulphonyl fluoride (PMSF), 1 mm DTT, 1 mmε-aminocaproic acid, pH 7.2]. The tissue was homogenized in an ice-cooled Potter-Elvehjem tissue grinder by a motor-driven teflon pestle (1000 rpm, 2 min) in 5 mL/g homogenization buffer. The homogenate was subsequently subfractionated in consecutive centrifugation steps at 4°C: debris and nuclei were removed by two rounds of centrifugation at 100 ×g, 10 min and 600 ×g, 10 min. The resulting post nuclear supernatant was centrifuged at 1950 ×g, 15 min to remove the heavy mitochondrial (HM) pellet. The supernatant was aspirated and centrifuged anew at 25 500 ×g, 20 min. From the pellet a fluffy grayish layer was carefully removed and the residual pellet was resuspended in an appropriate volume of ice-cold HB (1 mL/g liver) using a glass rod, recentrifuged and resuspended as given before. The suspension thus obtained is enriched in peroxisomes (crude peroxisomal preparation) and corresponds to the light mitochondrial (LM) fraction. To further purify peroxisomes from the LM fraction, OptiPrep (Axis-Shield) density gradients of sigmoidal shape with a density range of 1.12–1.26 g/mL were preformed, and 5 mL of the crude peroxisomal preparation were top-loaded each. Gradients were spun at an integrated force of 1252 × 106×g min, corresponding to a maximal relative centrifugal force of 39 000 ×g and a total centrifugation time of 45 min, using a Beckmann VTi50 vertical-type rotor. Highly purified peroxisomes banding at a density of 1.23 g/mL were carefully collected by aspiration and stored at −80°C until further use.

To obtain a cytosolic fraction, the supernatant resulting from the 25 500 ×g centrifugation step (see preparation of the LM fraction) was centrifuged for 30 min at 100 000 ×g. The supernatant was carefully aspirated with a glass pipette to avoid intake of the lipid phase banding at the top of the centrifugation tube and stored at −80°C.

For the isolation of mitochondria, OptiPrep centrifugation medium was used according to a protocol given by the manufacturer. For the preparation, the pellet of the 1950 ×g centrifugation corresponding to the HM fraction was washed with ice-cold HB and centrifuged for another time. The final pellet was resuspended in 25% OptiPrep, 137 mm sucrose, 1 mm EGTA, 20 mm Tricine/NaOH pH 7.8 and adjusted with a 50% OptiPrep solution to a final concentration of 25% by determination of the suspension's refractive index. This HM suspension was top-loaded onto a 27.5% OptiPrep solution (128 mm sucrose, 1 mm EGTA, 20 mm Tricine/NaOH pH 7.8) and overlayed with a 20% OptiPrep solution (156 mm sucrose, 1 mm EGTA, 20 mm Tricine/NaOH pH 7.8). This step gradient was centrifuged for 2 h at 50 000 ×g in an SW32 rotor (Beckmann). During centrifugation, peroxisomal contaminants form a pellet at the bottom of the tube, whereas microsomes float to the top of the gradient. The remainder of the gradient, containing purified mitochondria, was collected and stored at −80°C.

To enrich for the ER, the microsomal pellet, obtained by centrifuging the supernatant resulting from the 25 500 ×g centrifugation step for 30 min at 100 000 ×g, was resuspended in ice-cold HB and top-loaded onto a continuous 0.25–2 m sucrose gradient (5 mm MOPS, 1 mm EDTA, 0.1% ethanol, 2 mm PMSF, 1 mm DTT, 1 mmε-aminocaproic acid, pH 7.2). Centrifugation was carried out at 135 000 ×g for 180 min in a vertical-type Beckmann VTi50 rotor. Four individual bands were aspirated with a syringe and characterized by immunoblotting against organellar marker proteins.

For preparation of lysosomes, the total 25 000 ×g pellet was suspended in a 20% OptiPrep solution (156 mm sucrose, 1 mm EGTA, 20 mm Tricine/NaOH pH 7.8) and adjusted with a 50% OptiPrep solution to a final concentration of 20%. This self-generated gradient was centrifuged for 90 min using a Beckmann VTi50 vertical-type rotor. Fractions of 2 mL were collected from top to bottom and screened for β-glucuronidase activity. The fraction with the highest enzymatic activity was used as lysosomal fraction.

Nuclei were prepared from the 600 ×g pellet as described by Rickwood (43) using the method for aqueous nuclear preparations.

Mass spectrometry and immunoblotting

MS was accomplished as previously described (17) In short, 100 μg of each peroxisomal subfraction (matrix, integral membrane, peripheral membrane) was applied to MS. After trypsin digestion of the unseparated proteins, resulting peptides were fractionated using a two-dimensional liquid chromatography (first column: polysulfoethyl A, second column: C18). One hundred and ninety-two individual fractions were spotted on the ABI metal target and analyzed on an ABI-4700 proteomics analyzer. MS/MS-spectra were searched against rat and mice databases (Celera Discovery System, CDS) using GPS explorer and Mascot. For each MS/MS spectrum, a single peptide annotation with the highest Mascot score was retrieved. CDS protein sequence redundancy was removed by clustering the precursor protein sequences of the retrieved peptides using the cluster algorithm Cd hit (44). Immunoblots were performed according to the semi-dry method using standard procedures. Antibodies were purchased from Abcam (SOD1, ERp29, lamin 2) and BD Bioscience (ATP synthase, Pex5) or raised in rabbits as published previously (13) (catalase, LDH); an antibody specific to rat CCS was produced by immunization of guinea pigs with the peptide WEERGRPIAGQGRKDS coupled to KLH as published previously (20). After bleeding, the immunoglobulin (Ig) fraction of the serum was recovered by (NH4)2SO4 precipitation. An aliquot of the anti-CCS serum was incubated with 50 μm of the peptide used for immunization at 4°C overnight. Untreated serum and the peptide-depleted serum were subjected to centrifugation at 125 000 ×g for 90 min to remove large protein complexes, and supernatants were used for Western blots and for immunoelectron microscopy.

Immunoelectron microscopy

Rat liver was fixed by perfusion with 1.25% depolymerized paraformaldehyde, 0.5% glutaraldehyde, 0.025% picric acid in 0.05 m cacodylate buffer, pH 7.4. Dissected liver fragments were postfixed in the same fixative overnight and embedded in Epon according to standard protocols. Ultrathin sections were cut using a Reichert ultracut microtome (Reichert) and collected on polyvinyl-coated nickel grids. For immunolabeling of proteins, sections were washed in TTBS (20 mm Tris, 150 mm NaCl, 0.1% Tween-20 (v/v), 0.1% BSA (w/v) and 20 mm NaN3, pH 8.2) and incubated with polyclonal antibodies against rat liver catalase or against SOD1 (Abcam) raised in rabbit or against CCS raised in guinea pig (this study) in TTBS overnight at 4°C. After repeated washing with TTBS, sections were incubated with either gold-labeled (12 nm) goat antibodies against rabbit IgG (British BioCell) or with gold-labeled (10 nm) antibodies against guinea pig IgG (Aurion). Subsequently, sections were washed in water and were contrasted with uranyl acetate and lead citrate according to standard protocols. For double labeling of CCS and catalase, ultrathin sections of rat liver were incubated sequentially with CCS antiserum raised in guinea pig and gold-labeled (10 nm) anti-guinea pig IgG, followed by incubation with anticatalase raised in rabbit and gold-labeled (5 nm) anti-rabbit IgG according to the protocol described earlier. The same incubations were carried out using peptide-depleted anti-CCS antiserum (described above) to examine the specificity of the peroxisomal CCS labeling. For analysis of the subcellular distribution of CCS, 117 images (magnification 10 000×) covering about 450 μm2 of hepatocyte section were analyzed. Squares of 500 nm2 were placed on image areas identified as nucleus, mitochondria, rER or cytoplasm based on ultrastructural characteristics. For identification of peroxisomes, the specific labeling of the marker enzyme catalase with 5 nm gold particles was used. Labeling densities for CCS-IgG-gold immune complexes were calculated and expressed as gold particles per square micrometers. For verification of the specificity of peroxisomal CCS labeling, the CCS antiserum was compared with peptide-depleted antiserum (see above). Images of 65 catalase-positive peroxisomes each were photographed in catalase/CCS double-labeled sections, and the CCS labeling density per peroxisome was calculated. Specimens were viewed and captured on negative films using a Zeiss EM10 electron microscope (Zeiss).

Construction of plasmids and source of experimental reporter cell line

To construct myc- and flag-tagged expression plasmids for rat SOD1 and CCS, total RNA from rat liver was isolated using the RNeasy Mini Kit (Qiagen), according to the manufacturer's protocol. Rat mRNA sequences of SOD1 and CCS were reverse transcribed into cDNA by Superscript III reverse transcriptase (Invitrogen) using oligo-dT priming. The cDNA of rSOD1 and rCCS was amplified by primers flanking the open reading frame using Pfx polymerase (Invitrogen) and cloned into TOPO II vectors (Invitrogen). The correct open reading frame of both sequences was verified by sequencing at GATC. Myc- and flag-tagged variants of rat CCS, rat CCSE3– and rat SOD1 were constructed by standard DNA technologies using the pCMV-tag vector system (Stratgene). Deletion constructs of CCS were obtained by digestion of N-myc-CCS with PpuMI or PpuMI/EcoRI, respectively, and subsequent religation. The construction of a CHO line stably expressing the peroxisomal marker GFP-SKL (BGL69) was previously described (17).

Immunofluorescence

For transient transfections, modified CHO cells were plated on 24-well plate glass coverslips and transfected using the Gene Porter-system (Peqlab), according to the manufacturer's recommendations. Immunostaining was performed as described previously (45). Monoclonal antibodies against the flag and the myc epitope were obtained from Sigma and Abcam, respectively. For visualization of immune complexes, a Cy3-labeled goat anti-mouse antibody (Dianova) was used. Microscopical examinations were carried out on an Olympus BX50WI laser scanning microscope using a PlanApo 60×/oil objective with a numerical aperture of 1.40. Digital images were captured with the Fluoview software (version 2.1) and processed with Adobe Photoshop 7.0 and Adobe Illustrator 10.

Blue Native PAGE

Blue native gels were prepared as described by Schägger and von Jagow (46). After electrophoresis, the gel was cut into stripes and equilibrated for SDS–PAGE. For BN-PAGE, polyacrylamide gels sized 19 × 15 cm were cast with a 5–18% separating gel overlayed by a 4% stacking gel. The solutions used were composed of: (i) gel buffer 3×: 1.5 mε-aminocaproic acid, 150 mm Bis-Tris, pH 7.0; (ii) AB mix: 49.5% T (total concentrations of acrylamide and bisacrylamide monomers in grams per 100 mL) and 3% C (percentage by weight of the cross-linker relative to the monomer). The 5% polyacrylamide solution was composed of 1.8 mL AB mix, 6 mL 3× gel buffer, 100 μL ammonium persulfate (APS) (10%), 10 μL tetramethylethylendiamine (TEMED) and 10 mL H2O, the 15% solution of 5.5 mL AB mix, 5 mL 3× gel buffer, 50 μL APS (10%), 5 μL TEMED, 3 g glycerol supplemented with H2O to total volume of 15 mL. The stacking gel was composed of 0.5 mL AB mix, 2 mL 3× gel buffer, 50 μL APS, 5 μL TEMED and 3.45 mL H2O. After polymerization, gels were stored overnight at 4°C. For each gel lane, 30 μg of peroxisomal matrix or 100 μg of cytosol was applied after dilution with BN-PAGE sample buffer (750 mm aminocaproic acid, 50 mm Bis-Tris/HCl, 0.5 mm EDTA, 5% glycerol, pH 7.0) in a ratio of 1:4. To estimate the MW of proteins in the first dimension, the Native Mark Unstained Protein Standard (Invitrogen) was used. Cathode and anode buffers used for the separation were composed of 50 mm Tricine, 15 mm Bis-Tris, 0.02% Coomassie G250, pH 7.0 and 50 mm Bis-Tris adjusted to pH 7.0 with HCl, respectively. Electrophoresis was carried out at 100 V in the stacking and 300 V in the separating gel. After native electrophoresis, the gel was cut into lanes and incubated for 2 min, 60°C in 1% SDS, 0.3% DTT, 25 mm Tris, 192 mm glycine, 0.01% bromphenolblue. Subsequently, the lanes were loaded horizontally onto 12.5% SDS–PA gels and electrophoresed at 40 mA (stacking gel) and 60 mA (separating gel). Immunoblots were performed on a semi-dry instrument using a discontinuous buffer system. To identify unspecifically labeled spots, blots solely incubated with secondary antibodies were performed as controls (data not shown).

Acknowledgments

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

We thank H. Mohr, K. Rummer, A. Cordes, G. Hofbauer and E. Völck-Badouin for technical assistance. We are grateful to H. D. Fahimi and J. Kirsch for their comments on the manuscript. This work was supported by a grant from German Federal Ministry for Education and Research (BMBF).

References

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Figure S1: Ultrastructural localization of proteins was performed by immunoelectron microscopy with specific antibodies to CCS and catalase (A) or SOD1 (B) on sections of rat liver. Immune complexes were visualized by gold particles [10 nm for CCS and 5 nm for catalase in (A) and 12 nm for SOD1 in (B)]. Arrowheads indicate the labeling for CCS in (A) and for SOD1 in (B) and the crystalloid core of peroxisomes is labeled by &ast;. All three proteins were localized within the peroxisomal matrix. Note that CCS was also present in mitochondria (M) and the cytoplasm and that the majority of SOD1 was localized in the cytoplasm, whereas a minor portion could be found in peroxisomes. The bar in (A) represents 500 nm. For double immune labeling in (A), sections were incubated with guinea pig antibodies against CCS at 4°C overnight and immunocomplexes were visualized with gold-labeled (10 nm) antibodies against guinea pig IgG at 37°C for 30 min. After washing, sections were subsequently incubated with rabbit antibodies against catalase at 4°C overnight and immunocomplexes were detected with gold-labeled (5 nm) antibodies against rabbit IgG at 37°C for 30 min.

Figure S2: Localization of CCS and SOD1 in Pex5-deficient cells. Murine Pex5 (–/–) fibroblasts were transiently transfected with a plasmid for expression of myc-CCS (A) or a mixture of plasmids for expression of myc-CCS and SOD1-flag (B). Recombinant proteins were visualized with antibodies against the myc epitope (A) or against the flag epitope (B). Myc-CCS and SOD1-flag were both localized within the cytoplasm of transgene-expressing Pex5-deficient cells (arrowheads). A punctate peroxisomal staining pattern that was found in wild-type cells could not be observed in any of the transiently transfected Pex5 (–/–) cells, indicating that Pex5 is required for import of CCS or CCS/SOD1 complexes. Pex5 (–/–) fibroblasts were a gift of M. Baes (Leuven, Belgium). The bar in (B) represents 20 μm.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
TRA_966_sm_FigS1.tif2936KSupporting info item
TRA_966_sm_FigS2.tif4030KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.