Peroxisomal Membrane Protein Import Does Not Require Pex17p

Of the approximately 20 proteins required for peroxisome biogenesis, only four have been implicated in the process of peroxisomal membrane protein (PMP) import: Pex3p, Pex16p, Pex17p, and Pex19p. To improve our understanding of the role the Pex17p plays in PMP import, we examined the behavior of PMPs in a P. pastoris pex17 mutant. Relative to wild-type cells, pex17 cells appeared to have a mild reduction in PMP stability and slightly aberrant PMP behavior in subcellular fractionation experiments. However, we also found that the behavior of PMPs in the pex17 mutant was indistinguishable from PMP behavior in a pex5 mutant, which has no defect in PMP import, and far different from PMP behavior in a pex3 mutant, which has a bona fide defect in PMP import. Furthermore, we find that a pex14 mutant, which has no defect in PMP import, lacks detectable levels of Pex17p. Based on these and other results, we propose that Pex17p acts primarily in the matrix protein import pathway and does not play an important role in PMP import.

membrane biogenesis and PMP import (22,23), though it seems to function differently in the yeast Yarrowia lipolyitca (24) and appears to be absent from the yeast Saccharomyces cerevisiae (25). The fourth peroxin implicated in PMP biogenesis is P. pastoris Pex17p (26). This peroxin was identified in the yeast P. pastoris in a screen for mutants that are defective in the import of a PMP-GFP fusion protein (26). Furthermore, the pex17∆ mutant was thought to import PMPs less efficiently than wild type cells. Based on these results, Snyder et al. (26) and Subramani et al. (14) have proposed that Pex17p plays a specific role in PMP import, with the matrix protein import defect of pex17∆ cells resulting indirectly from this defect in PMP biogenesis. These results are somewhat different from those reported for S. cerevisiae PEX17, which appears to participate only in matrix protein import (8,27).
To improve our understanding of PMP import we began an investigation into the role of PEX17 in this process. An analysis of pex17∆ cells revealed that they do have slightly reduced levels of PMPs and that PMPs displayed an aberrant fractionation behavior, at least as compared to wild-type (WT) cells. However, we also observed these changes in pex5∆ mutants, which are impaired in matrix protein import but unaffected in All differential centrifugation and density gradient centrifugation experiments were performed as previously described (29).

Enzyme Assays
Assays were performed on supernatant and pellet fractions of post-nuclear supernatants generated from oleic acid-induced pex17∆, pex5∆, and WT strains spun for 1 hr at 250kg.
Glyceraldehyde-3-phosphate dehydrogenase (GAPD) activity was determined at 25°C by increasing A 340 over 160 sec. Each fraction was diluted 30 fold into the reaction buffer (13mM sodium pyrophosphate, pH 8.5, 26mM sodium arsenate, 25µM NAD + , 3 mM DTT), and the reaction was started with the addition of DL-glyceraldehyde-3-phosphate to a concentration of 500µM. Succinate dehydrogenase (SDH) activity was determined by incubating 25µl of each lysate at 37°C in the presence of 50mMpotassium phosphate, pH 6.8, 0.1% p-iodonitrotetrazolene, and 50mM Na 2 succinic acid for 10 mins. The reactions were stopped by the addition of TCA to 5%. Color was developed by the addition of 1ml ethylene glycol monomethyl ether, and the A 440 was read. 7 room temperature, washed thoroughly (4X, 10min total) in ddH 2 O, treated with 0.5% sodium meta-periodate for 15 min at room temperature, washed twice in ddH 2 O, and placed into filtered 2% uranyl acetate overnight at room temperature (in the dark).
Blocks were then rapidly dehydrated through a graded series of ETOH (4 0 C), followed by 3 washes in 100% ETOH (15 min each), then 2 washes with propylene oxide, and

PMP abundance and distribution in P. pastoris pex17∆ cells
Previous studies have shown that defects in PMP import result in rapid turnover and low steady state of most PMPs (8,(15)(16)(17)23,32). To determine whether this was also true for cells lacking Pex17p, we examined the abundance of three integral PMPs in a P. pastoris pex17∆ strain: Pex3p (33), Pex10p (30), and Pex12p (31). To control for the specificity of any phenotypes we might detect, we also examined the phenotypes of pex3∆ cells, which have a well-established defect in PMP import (8,15,33), and pex5∆ cells, which are defective in matrix protein import but are not defective in PMP import (7)(8)(9). These strains were grown in YPD to mid-log phase, washed, and incubated in oleate-containing media for 18 h. Cells were then harvested, lysed, and total cellular protein was collected by TCA precipitation. Equal amounts of protein from each strain were separated by SDS-PAGE and blotted with anti-PMP antibodies. As previously reported for pex3∆ mutants of humans and S. cerevisiae, P. pastoris pex3∆ mutants had low steady state levels of PMPs ( Fig. 1). Steady state levels of Pex3p, Pex10p, and Pex12p were also somewhat lower in pex17∆ cells than in WT cells, though they were clearly higher than in the pex3∆ strain. The levels of these three PMPs in the pex5∆ cells were the same as in pex17∆ cells. This was somewhat surprising given that all prior studies of PMP biogenesis in pex5∆ mutants support the hypothesis that Pex5p has no role in PMP import (7)(8)(9)(10)(11)(12)(13)34).
To better understand the role of Pex17p in PMP import, we next used subcellular fractionation experiments in an attempt to assess the import of PMPs in pex17∆ cells, as well as in pex3∆, pex5∆, and WT controls. Each strain was grown in YPD, incubated in oleate-containing medium for 18 hr, and converted to spheroplasts. Cells were lysed in a dounce homogenizer and cleared of nuclei and cell debris by low speed centrifugation.
The resulting postnuclear supernatants (PNSs) were separated by centrifugation at 25kg for 30 min. Equal proportions of each fraction were then separated by SDS-PAGE and blotted with antibodies to Pex3p, Pex10p, and Pex12p (Fig. 2). The levels of these PMPs in extracts prepared from pex3∆ cells were below the level of detection. In cells lacking Pex17p, ~20-30% of the Pex10p and the Pex12p and ~50% of the Pex3p were detected in the 25kg supernatant, which could be interpreted as evidence that pex17∆ cells have a partial PMP import defect. However, pex5∆ cells have a similar phenotype, raising questions about whether this assay is an accurate measure of PMP import.

Cells lacking Pex17p do not accumulate a pool of soluble, cytoplasmic PMPs
Snyder et al. (26) previously reported that the PMPs present in the 25kg supernatant fraction of pex17∆ mutant extracts were largely resistant to sedimentation at higher speeds (100Kg). Based in part on this result, they concluded that pex17∆ cells have a significant pool of soluble, cytosolic PMPs (26). However, other studies have suggested that cell homogenates contain significant levels of peroxisome-derived 'microsomes' that may pellet only at higher speeds (35-38). Post-nuclear supernatants were again generated from oleic acid-induced pex17∆, pex5∆, and WT strains and spun for 1 hr at 250kg.
Enzymatic assays of GAPD and SDH, cytosolic and mitochondrial enzymes respectively, demonstrate that these conditions are sufficient to pellet organelles but not soluble proteins. Equal proportions of each fraction were then separated by SDS-PAGE, by guest on March 24, 2020 http://www.jbc.org/ Downloaded from transferred to membranes and probed with antibodies to Pex3p, Pex10p, and Pex12p ( Fig. 3A). Under these conditions, all three PMPs were found primarily in the pellet fraction of pex17∆ cells. Similar results were observed for the pex5∆ mutant. Thus, it appears that PMPs present in the 25kg supernatant of pex17∆ and pex5∆ mutants do not represent pools of soluble, cytosolic PMPs.
The sedimentation of Pex3p, Pex10p, and Pex12p from pex17∆ and pex5∆ lysates could reflect the insertion of these proteins into peroxisomal membranes. However, their sedimentation behavior could also reflect a more complicated situation. For example, a portion of each PMP may be properly inserted into peroxisome membranes, while other subsets of these PMPs may only be peripherally associated with membranes or may exist in large protein aggregates. To determine the proportion of each PMP that was inserted into the peroxisome membrane in each cell type, we incubated the PNSs from pex17∆, pex5∆, and WT strains with 0.1 M Na 2 CO 3 , pH 11.5 (39), to extract non-integral proteins from cellular membranes. Integral membrane proteins were then collected in the pellet fraction by centrifugation at 250kg for 1 hr. Equal proportions of each fraction were then processed for immunoblot using antibodies to Pex3p, Pex10p and Pex12p (Fig. 3B). In both the pex17∆ and pex5∆ mutants, ~50% of their PMPs were released to the supernatant by carbonate extraction. The amounts of PMPs released from WT peroxisome membranes were similar to the amounts of PMPs released from pex17∆ and pex5∆ mutants, though they corresponded to only 10% of the total PMPs in WT cells, as opposed to 50% of the PMPs in these two mutants.

The P. pastoris pex14 mutant lacks both Pex14p and Pex17p
As part of our basic characterization of Pex17p, we examined its abundance in an array of P. pastoris pex mutants. We generated antibodies to recombinant Pex17p. These antibodies recognize a protein of ~25 kDa (the predicted molecular mass of Pex17p is 30,497 Da) in P. pastoris cell extracts. The protein recognized by these antibodies is absent from pex17∆ cells, colocalizes with peroxisomes in density gradient centrifugation experiments, and behaves as an integral PMP, indicating that these antibodies are specific for Pex17p (Fig. 4). These antibodies were then used to assess the abundance of Pex17p in the pex1, pex2, pex3, pex4, pex5, pex6, pex8, pex10, pex12, pex14, pex17, and pex22 mutants (Fig. 5). Whole cell lysates were generated by alkaline lysis, and equal amounts of protein from each strain were separated by SDS-PAGE, transferred to membranes, and probed with anti-Pex17p antibodies. As a control, we determined the levels of Pex13p in these same samples. As expected, Pex17p was absent from the pex17∆ mutant. Levels were also reduced in the pex3∆ mutant, consistent with the rapid degradation of all known PMPs in these cells (8,15,32) (see Fig. 1). However, we observed that Pex17p was also undetectable in the P. pastoris pex14 mutant, and it is interesting to note that the

Cells lacking Pex17p contain peroxisomes that resemble peroxisomes of pex5∆ cells
The hypothesis that Pex17p may not play a role in PMP import predicts that pex17∆ cells would display peroxisome morphologies that are largely indistinguishable from those of pex5∆ and pex14∆ cells. Electron micrographic examination of these mutants supports this hypothesis (Fig 6). After induction of peroxisomes by growth on methanol, wild type cells accumulate many large, electron-dense peroxisomes. In contrast, the pex3∆ mutant, that has a bona fide PMP import defect, lacks detectable peroxisomal structures.
Peroxisomal structures of similar abundance and morphology can be detected in the pex17∆, pex5∆, and pex14∆ mutants. These are much smaller and contain less electrondense material than WT cells, consistent with the matrix protein import defects in these pex mutants.

Discussion
We investigated the hypothesis that Pex17p plays an important role in PMP import (14,26). The first consideration in trying to determine whether a particular peroxin participates in matrix protein import or PMP import is to determine whether the peroxin is essential for PMP import. In the present study we established that P. pastoris pex17∆ cells are able to import PMPs and assemble peroxisome membranes, demonstrating conclusively that Pex17p is not required for PMP import. This aspect of the pex17∆ phenotype contrasts sharply with that of a bona fide PMP import mutant, pex3∆, in which PMPs are degraded rapidly due their PMP import defect (8,15,32) and are virtually undetectable.
The second consideration in assessing the role of Pex17p in PMP import is to determine whether its loss has a partial yet specific effect on PMP import. We did observe a slight reduction in PMP abundance and the proportion of PMPs that were inserted into the peroxisome membrane in P. pastoris pex17∆ cells. However, the question is whether any mutant that is defective in peroxisomal matrix enzyme import will also display these phenotypes, or whether they reflect a specific role for Pex17p in PMP import. We observed that cells lacking Pex5p also displayed these phenotypes.
Pex5p is the import receptor for PTS1-containing peroxisomal matrix enzymes and several previous studies have concluded that Pex5p does not play any role in PMP import (7)(8)(9)(10)(11)(12)(13)34). Therefore, it is reasonable to conclude from these data that Pex17p is unlikely to play a specific role in PMP import.
The phenotypes of the P. pastoris pex14 mutant (40) contribute independent evidence that P. pastoris Pex17p does not play a specific role in PMP import. Pex14p  (40), have reached the conclusion that they have no detectable defect in PMP import. We found that Pex17p levels are below the limit of detection in pex14 cells, demonstrating that the pex14 mutant is the equivalent of a pex14, pex17 double mutant. Since the P. pastoris pex14 mutant does not display any detectable defect in PMP import (40), it is highly unlikely that the P. pastoris pex17∆ mutant could have a PMP import defect.
The final argument against a role for Pex17p in PMP import comes from studies of S. cerevisiae Pex17p (27). Peroxisomal matrix protein import is severely affected in S. cerevisiae pex17∆ mutants but PMP import appears to be normal in these cells. In a systematic side-by-side analysis of S. cerevisiae pex mutants, Hettema et al. (8) established that PMP abundance and distribution in pex17∆ cells was the same as it was in the pex1, pex2, pex4, pex5, pex6, pex7, pex8, pex10, pex11, pex12, pex13, pex14, and pex15 mutants. Although there are some peroxins which function differently in different species, these are rare and there is no precedent for such differences between P. pastoris and S. cerevisiae peroxins.
In addition to casting doubt on the hypothesis that Pex17p plays an important role in PMP import (14,26), our data revealed that the fractionation behavior of PMPs in WT cells may be quite different from that observed in pex mutants that have no specific defect in PMP import. The pool of PMPs seen in pex5∆ and pex17∆ mutants that did not sediment at 25kg, which is sufficient to pellet most peroxisomes of WT cells (29) (see Fig. 2), was also seen in other matrix protein import mutants, including the pex4∆, pex10∆, and pex14 mutants (data not shown). One explanation for these results could be that the extremely small size of peroxisomes in pex5∆, pex17∆ and many other pex mutants precludes their quantitative sedimentation at speeds that are known to sediment the large, enzyme-rich peroxisomes of WT cells. After all, sedimentation behavior in these experiments is primarily a function of size rather than density, with larger structures being pelleted at lower centrifugation speeds and shorter centrifugation times. As for the observation that a higher proportion of PMPs can be extracted by high pH buffer in these mutants (~50%) as compared to WT cells (~10%), there is no difference between these mutants and WT cells in regard to the absolute amounts of PMPs extracted by high pH treatment (see Fig. 3B). The high pH-extractable PMPs we detected in both WT cells and pex mutants may include PMPs that are bound to the membrane but not yet inserted in the bilayers or merely large protein aggregates.
Although the above arguments indicate that there may, in fact, be no actual defect in PMP import in the pex5∆, pex17∆ and other pex mutants, it is also possible that any mutant that is defective in matrix enzyme import will display an ancillary, non-specific impairment in PMP import. Peroxisomes in these pex mutants are metabolically inactive, far smaller than normal peroxisomes, and appear to have far less surface area than peroxisomes of WT cells. This situation may delay PMP import, resulting in the accumulation of PMPs on the membrane surface and/or in protein aggregates, which could also explain our fractionation data. Regardless of the actual reason for our observation, it is clear that an appropriate set of WT and mutant control strains is required to distinguish between specific and non-specific defects in PMP import. by guest on March 24, 2020 http://www.jbc.org/ Downloaded from While the simplest interpretation of our data is that Pex17p acts directly in matrix protein import and plays no specific role in PMP import, it is not possible to completely exclude the possibility that Pex17p plays a direct, minor role in PMP import. However, if the available data is to be interpreted as evidence for such a role, then one must also conclude that Pex5p and Pex14p participate directly in PMP import, a conclusion that contradicts a large body of evidence (7-13) (40-42,44).
In addition to showing that Pex17p is unlikely to play an important role in PMP import and most likely plays a direct role in peroxisomal matrix protein import, we    Pex17p is found predominantly in the pellet fraction even after carbonate extraction, indicating that it is an integral PMP. Figure 5. Pex17p is undetectable in pex14cells. Whole-cell protein extracts were generated from P. pastoris pex mutants. Equal amounts of protein from each sample were separated by SDS-PAGE and analyzed by western blot. Pex17p is absent from pex17∆ cells, found at low levels in pex3∆ cells, and is absent from pex14cells.
Antibodies against Pex13p were used as a loading control.