Structural and functional analysis of matrix protein import into peroxisomes

The post-translational translocation of newly synthesized peroxisomal enzymes across the peroxisomal membrane is a crucial step in the biogenesis of peroxisomes and is essential for many metabolic processes in eukaryotic cells. The ability to import folded and even oligomeric proteins is a hallmark of peroxisomal protein import and imposes challenges to the peroxisomal import machinery: depending on the metabolic status of the cell, it has to (1) adapt to cargoes of varying size and abundance, while (2) ensuring selectivity and (3) directionality towards the peroxisomal lumen. Newly synthesized proteins having a peroxisomal targeting signal (PTS) are first recognized by soluble receptors in the cytosol and then targeted to the peroxisomal membrane. Here, the receptor/cargo complex transitions into an insoluble state by diverse interactions with a membrane bound docking complex, consisting of Pex13, Pex14 and, in yeast, additionally Pex17 and Dyn2. To gain a comprehensive understanding of this “docking event” of receptor/cargo complexes at the peroxisomal membrane, the first part of this work focusses on the characterization of the Pex14/17 complex from S. cerevisiae by an integrative structural biology approach. This was achieved using a recombinantly co expressed and purified Pex14/17 complex, which was studied solubilized in detergent as well as reconstituted in lipid nanodiscs and liposomes. A circular MSP1D1Δ4 to 6 nanodisc was used for cryo electron microscopy (cryo EM) analysis of the reconstituted protein complex and allowed to produce a rod-like three dimensional (3D) density map of single particles with an average resolution of 8.5 Å. By subjecting the detergent solubilized Pex14/17 complex to native mass spectrometry (MS) analysis, a 3:1 stoichiometry was determined. Using the homo bifunctional crosslinker bis(sulfosuccinimidyl)suberate (BS3) and thorough MS based downstream analysis of crosslinked protein complexes, eight different inter molecular crosslinks between Pex14 and Pex17 were identified. The distance constraints deduced from the identified crosslinks exclusively allowed a parallel arrangement of single entities in the density map. Furthermore, the second predicted coiled coil domain of Pex14 was found to be essential for Pex14 homo oligomerization into trimers. Lastly, the membrane topology of the Pex14/17 complex was determined to a N\(ᵢₙ\)-C\(ₒᵤₜ\) orientation by negative stain EM and cryo electron tomography (cryo ET). Finally, an integrative structural model of the peroxisomal membrane docking complex was established which suggests that rod like Pex14/Pex17 complexes reach into the cytosol and catch receptor/cargo complexes via flexible C terminal receptor binding domains (BDs) in Pex14 to recruit them to the peroxisomal membrane for the following translocation step. In the second part of the thesis the proposed structural model of the Pex14/17 docking complex was extended by the integration of Dyn2. To this end, a native MS based in vitro peptide binding assay and AlphaFold 3 (AF) structure predictions were conducted. Interestingly, Dyn2 was shown to bind two copies of Pex17 via a so far uncharacterized QT motif, which is present between two predicted coiled coil domains of the protein. Lastly, the predicted structural model was tested with publicly available BS3 crosslinks, which were found in Pex14 complexes purified from solubilized S. cerevisiae membrane fractions. Notably, the model is in line with all the reported inter molecular crosslinks between Dyn2 and Pex14 or Pex17. Taken together, the observed in vitro and in silico experiments suggest that single Pex14/17 hetero tetramers can form in a Dyn2 dependent manner higher order complexes with a 6:2:2 stoichiometry. This process potentially improves the efficacy of the docking process by increased avidity effects under demanding metabolic conditions. The pivotal role of Pex14 homo oligomerization was studied in more detail in the third part of this thesis by subjecting differentially truncated soluble variants of Pex14 to a real time native MS workflow. By mixing equimolar solutions of a longer variant, consisting of amino acids 116 to 341, and a shorter variant, consisting of amino acids 119 to 341, the introduced mass difference allowed for the monitoring of newly formed mixed oligomeric species over time. The acquired native MS data was analyzed in a second step by an ordinary differential equation-based mathematical model. This analysis revealed that Pex14 dimers are assembled at a slower rate compared to trimers, whereas dimers disassemble faster compared to trimers. In consequence, the observed subunit exchange reactions suggest that Pex14 trimers are more stable compared to dimers and they are likely the favored functional entity. Additionally, mathematical modeling indicates that Pex14 monomers, compared to Pex14 dimers and trimers, are suppressed in the native MS analysis and thus underestimated in abundance. This finding emphasizes the need for adequately controlling response factors obtained for different protonated species in native MS analyses. The fourth aim of this thesis was to identify and characterize so far unknown proteins with a role in the peroxisomal import of PTS2 cargo proteins. This was achieved by an affinity purification (AP) MS approach using the PTS2 specific co receptor Pex18 from S. cerevisiae as a bait. Co purified candidate proteins were tested by a thorough analysis of publicly available interaction databases and the so far uncharacterized protein Yjr012c was selected for further analysis. YJR012C and the open reading frame (ORF) C6ORF266 from H. sapiens encode for orthologous proteins, which were reported in earlier high throughput studies as direct interactor of the PTS2 receptor protein Pex7 in S. cerevisiae and human cells. Yeast cells lacking YJR012C showed a substantial growth defect, when cultured in oleic acid containing liquid medium. Furthermore, the PTS2 cargo proteins Pot1 and Gpd1 were mislocalized to cytosolic fractions. Quantitative stable isotope labeling with amino acids in cell culture (SILAC) MS analysis showed that cellular Pex7 levels are largely decreased in these cells, which likely causes the observed peroxisomal PTS2 import defect. The reported interactions of C6ORF266 were validated by immunoprecipitation (IP) experiments in various human cell lines, in vitro native polyacrylamide gel electrophoresis (PAGE) and size exclusion chromatography (SEC) based interactions studies. Additionally, AF 2 predictions of heteromeric Yjr012c/C6ORF266 complexes gave detailed mechanistic insights: Yjr012c/C6ORF266 directly interacts with Pex7 via a novel interaction interface at the bottom face of Pex7, they stabilize the Pex7/PTS2 interaction via an N terminal amphipathic α helix and do not block the binding of the PTS2 co receptor. Of note, the last observation is required for the peroxisomal targeting of the hetero tetrameric pre import complexes. Interestingly, a so far uncharacterized short linear motif with the consensus sequence [K/R]KPWE, present in Yjr012c/C6ORF266, was found to be essential for the interaction with Pex7. This motif was also found in the extreme N terminal part of Pex13 orthologs, exclusively from species having a Yjr012c/C6ORF266 ortholog, which points to a functional dependency. Indeed, mutating this motif in Pex13 impaired functional PTS2 protein import in yeast and human cells. In summary, due to the discovered role of Yjr012c and C6ORF266 in peroxisome biogenesis, we renamed these proteins to Pex39. Furthermore, I characterized in detail the role of the [K/R]KPWE motif in Pex39 and Pex13 for coordinated handover of PTS2 (co)receptor/cargo complexes at the peroxisomal membrane for subsequent translocation into the lumen of peroxisomes. In the last part of this thesis, protein interactions of the PTS1 receptor protein Pex5 from yeast with the PTS1 cargo protein Pcs60 and Pex8, an intraperoxisomal organizer of the peroxisomal matrix protein import machinery, were studied in cooperation with the group of Matthias Wilmanns (EMBL Hamburg and University Medical Center Hamburg Eppendorf) using an integrative structural biology approach. First, recombinantly expressed and purified full length ScPex5 and a N terminally truncated variant of Pex5 only comprising the tetratricopeptide repeat (TPR) domain were individually mixed with ScPcs60 and subjected to photo crosslinking using sulfo succinimidyl 4-4’-azipentanoate (SDA) followed by the identification of crosslinked residues via MS. The obtained data showed several interactions between the highly flexible N terminal part of Pex5 and Pcs60 in a 1:1 stoichiometry. Notably, even though most inter molecular crosslinks were not observed in the C terminal TPR domain of Pex5, the well characterized PTS1/TPR domain interaction was essential for complex formation. Interestingly, the N terminally truncated variant of Pex5 also showed several inter molecular crosslinks with Pcs60 in a 1:1 complex. Collectively, the acquired crosslinking MS (CLMS) data indicate a bimodal interaction between Pex5 and Pcs60 and serves as a resource for ongoing structural modeling studies using high resolution X ray data and AlphaFold 3 based structural modeling. When Pex8/Pex5 complexes from Pichia pastoris were studied by the same experimental approach, it was found that the identified inter molecular crosslinks support a three dimensional density map, which was obtained by cryo EM and exhibit an average resolution of 4.5 Å but only partially covered the flexible N terminal domain (NTD) of Pex5. The integrative structural model suggests that Pex8 consists of 12 α helical HEAT (Huntingtin, elongation factor 3 (EF3), protein phosphatase 2A (PP2A), yeast kinase TOR1) repeats and binds the TPR domain of Pex5 via its PTS1 as well the N terminal amphipathic helices of Pex5 on its conserved concave surface. The interaction of Pex5’s intr