The peroxisomal protein import machinery displays a preference for monomeric substrates

Abstract

Peroxisomal matrix proteins are synthesized on cytosolic ribosomes and transported by the shuttling receptor PEX5 to the peroxisomal membrane docking/translocation machinery, where they are translocated into the organelle matrix. Under certain experimental conditions this protein import machinery has the remarkable capacity to accept already oligomerized proteins, a property that has heavily influenced current models on the mechanism of peroxisomal protein import. However, whether or not oligomeric proteins are really the best and most frequent clients of this machinery remain unclear. In this work, we present three lines of evidence suggesting that the peroxisomal import machinery displays a preference for monomeric proteins. First, in agreement with previous findings on catalase, we show that PEX5 binds newly synthesized (monomeric) acyl-CoA oxidase 1 (ACOX1) and urate oxidase (UOX), potently inhibiting their oligomerization. Second, in vitro import experiments suggest that monomeric ACOX1 and UOX are better peroxisomal import substrates than the corresponding oligomeric forms. Finally, we provide data strongly suggesting that although ACOX1 lacking a peroxisomal targeting signal can be imported into peroxisomes when co-expressed with ACOX1 containing its targeting signal, this import pathway is inefficient.

Keywords: PEX5; acyl-CoA oxidase; docking/translocation machinery; peroxisomes; protein import; urate oxidase.

Figure 1.

Newly synthesized ACOX1 dimerizes in vitro, a process inhibited by PEX5. (a) ACOX1 dimerizes in vitro. Upper panel: ACOX1 and HA-tagged ACOX1 (2HA-ACOX1) were synthesized individually (lanes 1 and 2) or co-synthesized in the absence (−) or presence (+) of 1 µM recombinant PEX5 (lanes 4 and 5, respectively) for 30 min and subjected to a 4 h chase. A mixture of the two proteins synthesized individually (lane 3) and the co-synthesis reactions (lanes 4 and 5) were subjected to immunoprecipitation (IP) using anti-HA antibody agarose beads (lanes 6–8, respectively). Note that all samples were made chemically identical before immunoprecipitation by adding recombinant PEX5. Lower panel: An identical experiment was performed using 2HA-ACOX1 and ACOX1-Flag. IVT, in vitro transcription/translation. (b) Sedimentation behaviour of in vitro synthesized ACOX1. ACOX1 synthesized for 30 min (panel I), and ACOX1 synthesized for 30 min and chased for 4 h in the absence (panels II and III) or presence of 1 µM PEX5 (panel IV) were loaded onto the top of sucrose gradients supplemented with 1 µM of either PEX5 (panels III and IV) or a control protein (panels I and II). After centrifugation, fractions were collected from the bottom of the gradients and subjected to SDS-PAGE/autoradiography. Ovalbumin (OA), bovine serum albumin (BSA) and immunoglobulins (IgGs) were used as sedimentation coefficient standards. Peroxisomal matrix proteins from mouse liver were also subjected to this analysis. A Coomassie-stained gel is shown (panel V). Carbamoyl phosphate synthetase (CPS), 2-enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase (EHHADH), acyl-CoA oxidase I subunits a, b and c (ACOX1a, ACOX1b, ACOX1c, respectively) and catalase (Cat) were identified by nano-HPLC-MALDI-MS/MS (data not shown). (c) Monomeric and dimeric ACOX1 present different susceptibilities to proteinase K. ³⁵S-mACOX1 and ³⁵S-dACOX1 isolated from a sucrose gradient were treated with increasing concentrations of proteinase K (PK) for 40 min on ice. After protease inactivation, samples were analysed by SDS-PAGE/autoradiography. Numbers to the left indicate the molecular weights of protein standards. Arrow heads indicate proteolysis fragments of ACOX1 (see main text). (d) Dimeric ³⁵S-ACOX1 and native/peroxisomal ACOX1 display the same proteolysis profile. ³⁵S-dACOX1 isolated from a sucrose gradient and native ACOX1 (from mouse liver purified peroxisomes) were subjected to protease treatment in the presence of Triton X-100 and subjected to SDS-PAGE/autoradiography (left panel) or western blotting using antibodies directed to the 53-kDa ACOX1 polypeptide (central panel). The same blot was reprobed with an antibody directed to the 21-kDa polypeptide of ACOX1 (right panel). F, front of the gel.

Figure 2.

mACOX1 is a better peroxisomal import substrate than dACOX1. (a) ³⁵S-mACOX1 and ³⁵S-dACOX1 isolated from a sucrose gradient were subjected to in vitro import reactions in the presence (+) or absence (−) of the indicated recombinant proteins. After incubation, one-half of each sample was treated with proteinase K, as indicated. The organelles were then isolated and analysed by SDS-PAGE/autoradiography. The autoradiograph (upper panel) and the corresponding Ponceau S-stained membrane (lower panel) are shown. I₁, I₂—5% of ³⁵S-mACOX1 and ³⁵S-dACOX1, respectively, used in the assays. The arrow head indicates the 51 kDa protease-resistant fragment of ³⁵S-dACOX1. (b) Import kinetic analyses of ³⁵S-mACOX1 and ³⁵S-dACOX1. The two import reactions (each containing 2.5 mg of PNS) were performed in the presence of recombinant PEX5. Aliquots of each import reaction (containing 500 µg of PNS) were withdrawn at the indicated time points, treated with proteinase K and analysed as above. Note that the amount of ³⁵S-dACOX1 used in this experiment was approximately twofold that of ³⁵S-mACOX1 to obtain similar substrate concentrations. Lanes I—5% of the radiolabelled proteins present in each aliquot.

Figure 3.

Behaviour of UOX in in vitro homo-oligomerization and import assays. (a) In vitro synthesized UOX oligomerizes in a process inhibited by PEX5. UOX and HA-tagged UOX (2HA-UOX) were synthesized individually (lanes 1 and 2) or co-synthesized in the absence or presence of 1 µM recombinant PEX5 (lanes 4 and 5, respectively) for 45 min, and subjected to a 4 h chase. A mixture of the two proteins synthesized individually (lane 3) and the co-synthesis reactions (lanes 4 and 5) were subjected to immunoprecipitation (IP) using anti-HA antibody agarose beads (lanes 6–8, respectively) and analysed as described in figure 1a. IVT, in vitro transcription/translation. (b) Sedimentation analyses of in vitro synthesized UOX. Radiolabelled UOX synthesized for 45 min (panel I) or UOX synthesized for 45 min and chased for 4 h (panel II) were subjected to sucrose gradient centrifugation analyses. The sedimentation positions of ovalbumin (OA), bovine serum albumin (BSA) and immunoglobulins (IgGs) are also shown. (c) Monomeric and tetrameric UOX display different susceptibilities to protease treatment. ³⁵S-mUOX and ³⁵S-tUOX isolated from a sucrose gradient were subjected to proteinase K (PK) treatment (400 µg ml⁻¹, final concentration) and aliquots were withdrawn at the indicated time points. Samples were processed as described in figure 1c. (d) mUOX is a better substrate for the peroxisomal protein import machinery than tUOX. ³⁵S-mUOX and ³⁵S-tUOX isolated from a sucrose gradient were subjected to in vitro import assays in the presence (+) or absence (−) of the indicated recombinant proteins. After incubation, one-half of each sample was treated with proteinase K, as indicated. Isolated organelles were processed and analysed by SDS-PAGE/autoradiography. The autoradiograph (upper panel) and the corresponding Ponceau S-stained membrane (lower panel) are shown. I₁, I₂—5% of ³⁵S-mUOX and ³⁵S-tUOX, respectively, used in the assays. The arrow heads indicate protease-resistant fragments of ³⁵S-tUOX.

Figure 4.

ACOX1 lacking a peroxisomal targeting signal is inefficiently imported into peroxisomes when co-expressed with ACOX1 containing a PTS1. (a) COS-7 cells were transfected with plasmids encoding HA-tagged ACOX1 (2HA-ACOX1; panel I), a C-terminally Flag-tagged ACOX1 (ACOX1-Flag; panels II and III), or a HA-tagged ACOX1 containing a nuclear targeting sequence (2HA-ACOX1–3NLS; panel IV). Two days post-transfection the cells were fixed, counterstained with DAPI, and processed for immunofluorescence using an anti-PEX14 antibody (to label peroxisomes) and an anti-HA (panels I and IV) or anti-Flag antibody (panels II and III). Profile plots of fluorescence intensity (in percentage of pixel intensity) along the white arrows shown in the merged panels are also provided: blue line, DAPI staining; green line, anti-PEX14 fluorescence; red line, anti-HA or anti-Flag fluorescence. Scale bar, 10 µm. (b) Expression levels of tagged ACOX1 proteins in COS-7 cells. Untransfected cells (lanes ‘—’) or cells transfected with individual plasmids encoding ACOX1-Flag, 2HA-ACOX1 or 2HA-ACOX1–3NLS were analysed by western blot using an antibody against the 53 kDa polypeptide of ACOX1. The arrow heads indicate the tagged ACOX1 proteins. (c) COS-7 cells were transfected with mixtures of the plasmids encoding ACOX1-Flag and 2HA-ACOX1 at ratios of 1 : 10 and 1 : 30. Note that, as the total amount of plasmid in each of these mixtures was adjusted to 1 µg, ‘1’ corresponds to 10% and 3.3%, respectively, of the amount of plasmid DNA used in (a) and (b). The subcellular localization of each protein was then analysed by immunofluorescence 1, 2 and 3 days post-transfection (dpt). Cells in which ACOX1-Flag displays an exclusive cytosolic localization (Cyt; see upper panel for a representative example), a dual peroxisomal/cytosolic localization (PO/Cyt; middle panel), or an exclusive peroxisomal localization (PO; lower panel) were counted and expressed as percentage of ACOX1-Flag-expressing cells in the bar graph. (d) Exactly the same co-transfection strategy was used with plasmids encoding 2HA-ACOX1–3NLS and ACOX1-Flag. Cells in which ACOX1-Flag displays an exclusive cytosolic localization (Cyt; see upper panel for a representative example), a dual nuclear/cytosolic localization (Nuc/Cyt; middle panel) or an exclusive nuclear localization (Nuc; lower panel) were counted and expressed as percentage of ACOX1-Flag-expressing cells in the bar graph. Note that at least 200 cells were analysed per condition. Scale bar, 10 µm.

Figure 5.

Working model for the initial steps of the peroxisomal matrix protein import pathway. Newly synthesized proteins are bound by chaperones during or immediately after their synthesis on cytosolic ribosomes. Successfully folded proteins are then released as near-native monomers and bound by cytosolic PEX5. These monomers are transported to the peroxisomal matrix where oligomerization can occur. Peroxisomal matrix proteins that do not fold correctly (e.g. as a result of mutation) are not released from the chaperones. Retained proteins may be degraded or mistargeted, as is the case of some mutant forms of alanine-glyoxylate amino transferase and 2-enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase which target mitochondria [75,76]. Some proteins may also be imported after oligomerization. This pathway may be particularly used when matrix proteins are overexpressed and the import machinery becomes rate-limiting. Note that this pathway is rather inefficient for the two proteins characterized in this work and cannot be used for all those proteins that no longer expose their PTS upon oligomerization.