A PEX5 missense allele preferentially disrupts PTS1 cargo import into Arabidopsis peroxisomes

Abstract The sorting of eukaryotic proteins to various organellar destinations requires receptors that recognize cargo protein targeting signals and facilitate transport into the organelle. One such receptor is the peroxin PEX5, which recruits cytosolic cargo carrying a peroxisome‐targeting signal (PTS) type 1 (PTS1) for delivery into the peroxisomal lumen (matrix). In plants and mammals, PEX5 is also indirectly required for peroxisomal import of proteins carrying a PTS2 signal because PEX5 binds the PTS2 receptor, bringing the associated PTS2 cargo to the peroxisome along with PTS1 cargo. Despite PEX5 being the PTS1 cargo receptor, previously identified Arabidopsis pex5 mutants display either impairment of both PTS1 and PTS2 import or defects only in PTS2 import. Here, we report the first Arabidopsis pex5 mutant with an exclusive PTS1 import defect. In addition to markedly diminished GFP‐PTS1 import and decreased pex5‐2 protein accumulation, this pex5‐2 mutant shows typical peroxisome‐related defects, including inefficient β‐oxidation and reduced growth. Growth at reduced or elevated temperatures ameliorated or exacerbated pex5‐2 peroxisome‐related defects, respectively, without markedly changing pex5‐2 protein levels. In contrast to the diminished PTS1 import, PTS2 processing was only slightly impaired and PTS2‐GFP import appeared normal in pex5‐2. This finding suggests that even minor peroxisomal localization of the PTS1 protein DEG15, the PTS2‐processing protease, is sufficient to maintain robust PTS2 processing.

Analysis of mutants defective in peroxisome cargo receptors can provide insight into the import machinery. Only two Arabidopsis pex5 mutants, pex5-10 and pex5-1, have been described. pex5-10 carries a T-DNA insertion in the fifth exon of PEX5  that results in the skipping of this exon and production of an internally deleted pex5-10 protein lacking several predicted PEX14-binding motifs ( Figure 1a) (Khan & Zolman, 2010). The pex5-10 mutant, like pex5 RNAi lines (Hayashi et al., 2005), has defects in both PTS1 and PTS2 import (Khan & Zolman, 2010;. pex5-1 is a missense allele that creates a Ser318Leu substitution (Zolman et al., 2000) in the predicted PEX7-binding domain (Figure 1a), and the pex5-1 mutant has impaired PTS2 import but wild-type PTS1 import (Woodward & Bartel, 2005a). Similarly, Arabidopsis pex7 mutants and RNAi lines display defects in PTS2 import (Hayashi et al., 2005;Ramón & Bartel, 2010;Woodward & Bartel, 2005a). In addition to PTS2 import defects, Arabidopsis pex7 mutants show decreased PEX5 levels and defects in PTS1 import (Ramón & Bartel, 2010), indicating that PEX5 and PEX7 are interdependent. As Arabidopsis pex5 mutants with exclusively PTS1 import defects have not been reported, distinguishing the functions of PTS1 and PTS2 import in plants has been challenging.
In this study, we describe a novel pex5 missense mutation (pex5-2) recovered from a forward genetic screen for β-oxidation defects.
The pex5-2 mutant exhibited reduced growth, low PEX5 levels, and decreased peroxisomal import of GFP-PTS1 protein. In contrast, pex5-2 displayed robust PTS2-GFP import and only slight defects in PTS2 protein processing, suggesting that relatively little PTS1 import may be sufficient to efficiently cleave PTS2 signals. Some pex5-2 deficiencies were exacerbated at elevated growth temperature and ameliorated at lowered growth temperature, suggesting that PEX5 function and/or pex5-2 dysfunction is impacted by temperature. The distinct and overlapping defects of the Arabidopsis pex5-1, pex5-2, and pex5-10 mutants will allow continued elucidation of the relationships between PTS1 and PTS2 import in plants.
Seeds were surface-sterilized with bleach solution (3% NaOCl and 0.01% Triton X-100), washed twice with sterile water, suspended in 0.1% (w/v) agar, and stratified in dark at 4°C for 1-3 days. Stratified seeds were plated on plant nutrient (PN) media (Haughn & Somerville, 1986) solidified with 0.6% (w/v) agar and supplemented with 0.5% sucrose (PNS) as indicated. PNS plates were supplemented with IBA (from a 100 mM IBA stock in ethanol) as indicated. Plates were incubated at 22°C in yellow-filtered light (Stasinopoulos & Hangarter, 1990) for the indicated number of days, and light-grown root lengths were measured and/or seedlings were used for immunoblotting or confocal microscopy. For temperature experiments, plates were incubated at 22°C in yellow-filtered light for 1 day and then incubated at 16, 22, or 28°C for seven additional days for light-grown experiments or wrapped in aluminum foil and grown at 16, 22, or 28°C for four additional days for dark-grown experiments. All experiments except the initial characterization (Supporting Information Figure S1) were repeated at least twice with similar results; representative results are shown.

| pex5-2 isolation
Ethyl methanesulfonate (EMS) mutagenesis of wild-type seeds carrying 35S:GFP-PTS1 was previously described (Rinaldi et al., 2016). M 2 seeds were grown for about 2 weeks in yellow-filtered light on PNS supplemented with 100 mM NaCl and 12 μM IBA, and putative mutants with elongated roots were transferred to soil for seed production. M 3 lines displaying resistance to 10 μM IBA (with or without 100 mM NaCl) and wild-type-like sensitivity to 80 nM 2,4-dichlorophenoxyacetic acid were retained for further analysis.

| Whole-genome sequencing
Approximately 100 pex5-2 M 3 seeds were plated on PNS covered with sterile filter paper. Genomic DNA was purified (Thole, Beisner, Liu, Venkova, & Strader, 2014) from 18-day-old light-grown seedlings and sequenced by the Genome Technology Access Center at Washington University in St. Louis. Sequence reads were aligned with the TAIR 10 build of the A. thaliana Col-0 genome using Novoalign (Novocraft; http://novocraft.com). SNPs were identified using SAMtools (Li, 2011;Li et al., 2009) and annotated with snpEFF (Cingolani et al., 2012). Mutations were filtered using a script prioritizing homozygous EMS-derived mutations (G-to-A and C-to-T) causing non-synonymous amino acid changes or altering splice sites. We disregarded mutations that were present in our laboratory stock of wild-type Col-0.
Identifiers of genes with homozygous EMS-consistent mutations are displayed in Supporting Information Figure S2.

| Statistical analysis
SPSS Statistics software (Version 24.0.0.0) was used to analyze measurements. One-way analysis of variance (ANOVA) followed by the Duncan's post hoc test was used to determine the significance of differences among genotypes or treatments.

| pex5-2 displays peroxisome-related defects that are rescued by PEX5 overexpression
As β-oxidation is exclusively peroxisomal in Arabidopsis (reviewed in Graham, 2008), we can use β-oxidation efficiency to assess peroxisome function (reviewed in . The predominant naturally occurring auxin, indole-3-acetic acid (IAA), plays critical roles in plant growth and development by modulating cell identity, division, and elongation (Woodward & Bartel, 2005b). One IAA precursor, indole-3-butyric acid (IBA), is converted into IAA via β-oxidation in peroxisomes Strader, Culler, Cohen, & Bartel, 2010). Thus, in seedlings with functioning peroxisomes, IBA treatment confers high-auxin phenotypes including reduced root and hypocotyl elongation Zolman et al., 2000), and IBA-resistance screens have uncovered genes needed for peroxisome biogenesis and function (reviewed in Bartel et al., 2014). As salt increases Arabidopsis peroxisome abundance (Fahy et al., 2017;Frick & Strader, 2018;Mitsuya et al., 2010), we reasoned that screening for IBA resistance in the presence of salt might uncover factors impacting salinity-induced peroxisome proliferation. We therefore identified seedlings with elongated roots on normally inhibitory concentrations of IBA and NaCl. Subsequent analyses of a mutant emerging from this screen showed reduced response to IBA in the presence of NaCl, but this mutant was even less IBA responsive in the absence of NaCl (Supporting Information Figure S1a), suggesting that we had not disrupted a proliferation-related gene.
Moreover, root growth of this mutant was also resistant to inhibition by 2,4-dichlorophenoxybutryic acid (Supporting Information Figure S1b), which, like IBA, requires peroxisomal chain shortening for activity (Hayashi, Toriyama, Kondo, & Nishimura, 1998). In contrast, the mutant root growth was inhibited similar to wild type by the synthetic auxin 2,4-dichlorophenoxyacetic acid (Supporting Information Figure S1b), suggesting that general auxin responsiveness was intact. We concluded that the mutant defects stemmed from decreased peroxisome function and we selected the mutant for in-depth analysis.
Whole-genome sequencing of genomic DNA from this mutant revealed a homozygous G-to-A missense mutation in the PEX5 gene (Supporting Information Figure S2), and we named the mutant pex5-2 ( Figure 1). The pex5-2 mutation results in a Gly498Arg substitution in one of the seven tetratricopeptide repeat (TPR) domains ( Figure 1) that comprise the PTS1-binding region of PEX5 (Gatto, Geisbrecht, Gould, & Berg, 2000;Terlecky, Nuttley, McCollum, Sock, & Subramani, 1995). The crystal structure of the human PEX5 TPR domain (Stanley et al., 2006) reveals that each TPR consists of two αhelices that pack together to form a PTS1-binding pocket (Figure 1c).
The Gly498Arg substitution in pex5-2 is in the middle of the first predicted α-helix of TPR2 (Figure 1b,c) and alters a Gly residue that is conserved in diverse PEX5 orthologs (Figure 1b).
In addition to strong IBA resistance (Figure 2a), the pex5-2 mutant exhibited a slight defect in processing of the peroxisomal malate dehydrogenase (PMDH) PTS2 protein (Figure 2b). IBA resistance was closely linked to the pex5-2 mutation; 12 of 12 IBA-resistant F 2 seedlings from a backcross to wild type were pex5-2 homozygotes.
These linkage and complementation experiments confirmed that the identified pex5-2 mutation was causing the observed peroxisomerelated impairments.
pex5-2 also displayed growth defects. Unlike pex5-1 shoots, which resembled wild type, pex5-2 shoots were smaller than wild type and more similar to pex5-10 ( Figure 2c). In contrast, pex5-2 seedling roots elongated similar to wild type and pex5-1 on sucrose-supplemented media (Figure 2a,c). As with IBA responsiveness (Figure 2a), expressing wild-type PEX5 in the pex5-2 mutant improved rosette size in the mutant (Figure 2c), indicating that these defects were caused by decreased PEX5 function.

| pex5-2 exhibits PTS1 but not PTS2 import defects
As the pex5-2 mutant presented more complete IBA resistance but less severe defects in processing PTS2 proteins than pex5-1, we directly compared peroxisomal matrix protein import in these alleles by using confocal microscopy to visualize import of PTS1and PTS2-tagged GFP reporters (Woodward & Bartel, 2005a;Zolman & Bartel, 2004). Wild type showed the expected GFP-PTS1 and PTS2-GFP puncta in seedling cotyledons and hypocotyls (Figure 3a,b), indicating efficient PTS1 and PTS2 import. As previously reported (Woodward & Bartel, 2005a) Figure S4). Quantification of punctate (peroxisomal) versus dispersed (cytosolic) fluorescence revealed that pex5-2 imported only a fraction (20%-50% in cotyledons; 5%-12% in hypocotyls) of GFP-PTS1, whereas wild type and pex5-1 imported more than 90% of GFP-PTS1 (Supporting Information Figure S3). Conversely, pex5-2 imported PTS2-GFP at least as well as wild type (over 90%) compared to less than 15% PTS2-GFP import for pex5-1 (Supporting Information Figure S4). Although PTS2-GFP fluorescence appeared brighter in pex5-2 than wild type (Figure 3b), images collected at different gain settings revealed that peroxisomes were of similar sizes in pex5-2 and wild type (Supporting Information Figure   S4a,b), and immunoblotting revealed more GFP in the pex5-2 line compared to wild type or pex5-1 (Supporting Information Figure S4c). As the transgene in pex5-2 was introduced by crossing from the wild-type line, this expression difference is likely due to different degrees of silencing in the different lines. We concluded that the pex5-2 lesion in the PTS1-binding TPR domain ( Figure 1) specifically impaired PTS1 import while leaving PTS2 import intact.

| Elevated growth temperature exacerbates pex5-2 physiological and molecular defects
After importing PTS1 cargo, PEX5 is ubiquitinated via the peroxisomal ubiquitination machinery. PEX4 is a ubiquitin-conjugating enzyme implicated in the ubiquitination that facilitates PEX5 retrotranslocation from the peroxisome membrane. The peroxisomal impairments of the pex4-1 mutant  are less severe when seedlings are grown at elevated temperature in the dark (Kao & Bartel, 2015). Moreover, PEX5 levels are lower in dark-grown seedlings at elevated growth temperature (Kao & Bartel, 2015). These results suggest the possibility that accumulated PEX5 protein contributes to pex4-1 physiological defects. To examine whether growth temperature also impacted pex5-2 phenotypes, we compared growth at three temperatures.
In light-grown seedlings, pex5-2 roots were highly IBA resistant when grown at all tested temperatures (16, 22, or 28°C) (Figure 4a), whereas pex5-2 root growth without sucrose was impaired at the normal growth temperature (22°C), further impaired at 28°C, but wild type-like at 16°C (Figure 4b). Similar to light-grown roots, pex5-2 dark-grown hypocotyls were highly IBA resistant regardless of growth temperature (Figure 4c), whereas pex5-2 hypocotyl growth without sucrose was most impaired at high temperature ( Figure 4d). Moreover, growth at 28°C resulted in slight accumulation of the precursor form of PMDH in light-grown pex5-2 seedlings (Figure 4e), suggesting worsened PTS2 processing at higher temperature, whereas PMDH processing was nearly complete in pex5-2 seedlings grown at 16°C (Figure 4e), suggesting improved PTS2 processing at low temperature. In contrast to pex5-2, PMDH processing in pex4-1 was improved at higher temperature and exacerbated at lower temperature ( Figure 4e). Temperature seemed to have a less severe impact on pex5-1 than on pex5-2. For example, PTS2 processing of PMDH was unchanged in pex5-1 grown at various temperatures ( Figure 4e).
We also examined PEX5 levels following growth at these three temperatures. Interestingly, the general decline in PEX5 levels that accompanied higher growth temperature in dark-grown wild-type seedlings (Figure 4f; Kao & Bartel, 2015) was not observed in lightgrown seedlings (Figure 4e). pex5-2 seedlings generally accumulated less PEX5 protein than wild type at all temperatures tested ( Figure 4e,f), suggesting that the worsened physiological ( Figure 4b) and PTS2-processing defects (Figure 4e) of pex5-2 at high temperature did not stem from a magnified decrease in overall pex5-2 protein level at high temperature.
F I G U R E 2 pex5-2 displays peroxisome-related defects that are rescued by PEX5 overexpression. (a) pex5-2 is resistant to the inhibitory effects of IBA on root elongation. Bars indicate mean root lengths of 8-day-old seedlings grown in light in the absence or presence of 10 μM IBA on media supplemented with 0.5% sucrose (n = 12). Error bars indicate standard deviations. Different letters within bars indicate significant differences between genotypes following IBA treatment (one-way ANOVA, p < 0.001). Percentages above bars indicate relative elongation for each line on IBA compared to mock conditions. (b) pex5-2 displays a slight defect in PTS2 processing. Extracts from 8-day-old light-grown seedlings were processed for immunoblotting and probed with antibodies to the indicated proteins. PMDH and thiolase are synthesized as precursor (p) proteins that are processed into mature forms (m) after peroxisome entry. HSC70 was used to monitor loading. (c) pex5-2 displays slight growth defects. Plants were grown on agar-based medium containing 0.5% sucrose for 8 or 15 days, and seedlings were moved to a new plate and photographed
F I G U R E 4 Growth at elevated temperature exacerbates some pex5-2 defects. (a-d) Seedlings were grown in the light (a, b) or the dark (c, d) on medium containing sucrose, containing sucrose and IBA (a, c), or lacking sucrose (b, d). Light-grown root lengths (a, b) or dark-grown hypocotyl lengths (c, d) were normalized to the corresponding mean length on sucrose-supplemented medium. Bars show the means of three (28°C) or four (16°C and 22°C) biological replicates (each with ≥12 seedlings); error bars show standard deviations. Different lowercase letters within bars indicate significant differences within a genotype at different temperatures (one-way ANOVA, p < 0.05). Different uppercase letters above bars indicate significant differences between genotypes at the same temperature (one-way ANOVA, p < 0.05). (e, f) Extracts from seedlings grown at the indicated temperatures in the light (e) or dark (f) were processed for immunoblotting and probed with antibodies to the indicated proteins. PEX5 levels (quantified using ImageJ) were normalized to HSC70 and then to the 22°C Wt level (set at 1.0) to give the numbers below the PEX5 panels (means of three biological replicates, including the one shown). PMDH and thiolase are synthesized as precursor (p) proteins that are processed into mature forms (m) after peroxisome entry. The percentages of unprocessed PMDH (means of three biological replicates, including the one shown) in the mutants are shown below the PMDH panel. HSC70 was used to monitor loading The Gly498Arg substitution in pex5-2 is in the second of seven TPR domains in the PEX5 C-terminal region (Figure 1). Although examination of the crystal structure of the human PEX5 TPR domain (Stanley et al., 2006) suggests that the affected Gly residue does not directly interact with the PTS1 of the cargo protein ( Figure 1c), this residue is conserved in PEX5 orthologs in metazoans, trypanosomes, and fungi (Figure 1b). PTS1 import was substantially impaired in the pex5-2 mutant (Figure 3a), suggesting that the Gly-to-Arg substitution might impede folding of TPR2 or the TPR domain in general, thus reducing PTS1 binding.

| Genetic and environmental controls of PEX5 levels and function
Growth temperature can influence peroxisome function in pex mutants. For example, reduced growth temperature improves import of catalase, a PTS1 protein, in mammalian cell lines carrying mutations found in several groups of peroxisome biogenesis disorder patients (Fujiwara et al., 2000;Imamura, Tamura, et al., 1998;. In contrast, growth at elevated temperature decreases PEX5 levels and improves growth, IBA responsiveness, and PTS2 processing in dark-grown pex4 seedlings ( Figure 4f; Kao & Bartel, 2015). The effects of temperature on pex4 defects suggest that the detrimental impact of excess PEX5 in the peroxisomal membrane when PEX4 is dysfunctional is partially relieved by decreased overall PEX5 levels at elevated temperature (Kao & Bartel, 2015).
Seedling pex5-2 protein levels were slightly lower than pex5-1 or wild-type PEX5 protein levels (Figure 4e,f). Decreased pex5-2 protein levels could reflect a TPR-folding defect that increases PEX5 susceptibility to degradation. Alternatively or in addition, PTS1 cargo binding, which we expect to be impaired by the pex5-2 mutation, might protect PEX5 from degradation. Low growth temperature slightly increased pex5-2 protein levels in dark-grown seedlings but not light-grown seedlings (Figure 4e,f). However, because we observed the most dramatic impacts of temperature on pex5-2 physiological and molecular defects in light-grown seedlings (Figure 4b,e), it seems likely that pex5-2 defects stem more directly from altered pex5-2 function caused by the Gly498Arg change and not solely from decreased pex5-2 levels.
It will be interesting to learn whether the apparent pex5-2 instability that we observe can be attributed to the peroxisomal ubiquitination machinery.
When PEX5 retrotranslocation is inefficient, as in mammalian pex1 mutants, PEX5 ubiquitination is associated with peroxisome degradation via specialized autophagy (Law et al., 2017). Similarly, genetically preventing autophagy improves peroxisome function in Arabidopsis pex1 and pex6 mutants Rinaldi et al., 2017). The allelic series of pex5 mutants may be useful in future dissection of the possible role of PEX5 in promoting autophagy of peroxisomes (pexophagy) in plants.

| Consequences and causes of PTS1-specific import defects
One hallmark of pex mutants is reduced PTS2 processing (reviewed in Bartel et al., 2014), which can vary in severity in different tissues or ages (Kao et al., 2016). The protease DEG15 cleaves the N-terminal PTS2-containing region (Helm et al., 2007;Schuhmann et al., 2008). As DEG15 is a PTS1 protein, PTS2 processing is expected to require robust PTS1 and PTS2 import. Interestingly, however, the strong PTS1 import defect observed in pex5-2 (Figure 3a) was accompanied by only minimal PTS2-processing defects (Figures 2b, 4e,f), suggesting that a small amount of DEG15 import is sufficient for substantial PTS2 processing and that more severe PTS2-processing defects probably reflect primarily PTS2 import defects. Of course, the ability to detect PTS2-processing defects relies on the stability of the precursor protein in the cytosol, which might vary for different proteins. For example, thiolase precursor levels increase in pex seedlings treated with the MG132 proteasome inhibitor (Kao & Bartel, 2015), implying that cytosolic thiolase is susceptible to ubiquitin-dependent degradation. Differences in cytosolic precursor stability could contribute to the apparent differences in PMDH and thiolase-processing efficiencies in pex5-2 (Figures 2b and 4e) and other pex mutants Gonzalez et al., 2017Gonzalez et al., , 2018Kao et al., 2016;Monroe-Augustus et al., 2011).
In addition to pex5-2, several other Arabidopsis pex mutants display impaired PTS1 import but apparently normal PTS2 import.
Salt (NaCl) increases Arabidopsis peroxisome abundance (Fahy et al., 2017;Frick & Strader, 2018;Mitsuya et al., 2010) and enhances the inhibitory effect of IBA (Supporting Information Figure S1a), presumably because increased numbers of peroxisomes host more β-oxidation. Although pex5-2 was selected in a screen for salinity-related peroxisome proliferation factors, subsequent testing revealed that salt still increased IBA responsiveness in pex5-2 (Supporting Information Figure S1a), suggesting that the peroxisome proliferation machinery remained functional and that PTS1 import is not necessary for this response. In contrast, salt did not similarly improve IBA responsiveness in pex6-1 roots (Supporting Information Figure S1a), hinting that PEX6 may be involved in salt-induced peroxisome proliferation. Future identification of mutants that do not increase IBA responsiveness in response to salt treatment might identify additional components of this response.

| CON CLUS IONS
The metabolic activities compartmentalized in Arabidopsis peroxisomes allow quantification of peroxisome function and dysfunction in an intact multicellular organism. The pex5-2 mutant described here provides insights into peroxisomal matrix protein import and the relationship between PTS1 import and PTS2 protein processing. With its PTS1-specific defects, the pex5-2 mutant extends the allelic series that includes the PTS2-specific pex5-1 (Woodward & Bartel, 2005a;Zolman et al., 2000) and the general import defective pex5-10 (Khan & Zolman, 2010;Zolman et al., 2005), providing a valuable asset for future peroxisome research.

ACK N OWLED G M ENTS
We thank Steven M. Smith for antibodies recognizing PMDH, Lucia C. Strader for assistance with whole-genome sequencing, Joseph Faust for developing the script to facilitate analysis of wholegenome sequencing data, Mauro A. Rinaldi for mutagenized seeds of wild type carrying 35S:GFP-PTS1, and Sammira Rouhani for assisting with the mutant screen. We are grateful for critical comments on the manuscript from Kathryn Smith, Melissa Traver, Zachary Wright, and Pierce Young. This research was supported by grants from the National Institutes of Health (NIH, R01GM079177), the National Science Foundation (MCB-1516966), and the Robert A.
Welch Foundation (C-1309) to BB. Confocal microscopy equipment was obtained through a Shared Instrumentation Grant (NIH S10RR026399). Genomic sequencing at the Genome Technology Access Center at Washington University School of Medicine was supported by the NIH (P30CA91842 and UL1RR024992). KJP was supported in part by a Spurlino Undergraduate Summer Fellowship, YTK was supported in part by the Studying Abroad Scholarship from the Taiwan Ministry of Education, and RJL was supported in part by the NIH (F31GM125367).

AUTH O R CO NTR I B UTI O N S
YTK and BB designed the research; YTK and KJP conducted the mutant screen; KJP performed the physiological and molecular characterizations; RJL, YTK, and KJP performed the microscopy; YTK, KJP, and BB wrote the manuscript; all authors revised the manuscript and approved the final version.