Peroxisome Function, Biogenesis, and Dynamics in Plants

Abstract

Recent advances highlight understanding of the diversity of peroxisome contributions to plant biology and the mechanisms through which these essential organelles are generated.

Eukaryotic cells employ organellar compartmentalization to increase efficiency of cellular processes and protect cellular components from harmful products, such as reactive oxygen species. Peroxisomes are organelles that sequester diverse oxidative reactions and play important roles in metabolism, reactive oxygen species detoxification, and signaling. Oxidative pathways housed in peroxisomes include fatty acid β-oxidation, which contributes to embryogenesis, seedling growth, and stomatal opening. Other peroxisomal enzymes enable photorespiration, which increases photosynthetic efficiency. Peroxisomes contribute to the synthesis of critical signaling molecules including the jasmonic acid, auxin, and salicylic acid phytohormones. Peroxisomes lack DNA; peroxisomal proteins are encoded in nuclear DNA and posttranslationally enter the organelle. Recent studies have begun to fill gaps in our understanding of how peroxisomal proteins are imported, regulated, and degraded. Despite this progress, much remains to be learned about how peroxisomes originate from the ER, divide, and are degraded through pexophagy, a form of organelle-specific autophagy. Peroxisomes play vital roles in multiple aspects of plant life, and in this review, we highlight recent advances in our understanding of plant peroxisome functions, biogenesis, and dynamics, while pointing out areas where additional studies are needed.

PEROXISOME FUNCTIONS

Fatty Acid β-Oxidation: Not Just for Seedlings

Peroxisomes house fatty acid β-oxidation, the process by which fatty acids are catabolized into acetyl-CoA (Fig. 1). Fatty acyl-CoA esters are transported into the peroxisome by the ATP-dependent transporter PEROXISOMAL ABC TRANSPORTER1 (PXA1)/COMATOSE/PEROXISOME DEFECTIVE3 (PED3) (Zolman et al., 2001; Footitt et al., 2002; Hayashi et al., 2002; Nyathi et al., 2010). Curiously, the fatty acyl-CoA ester is hydrolyzed during transport (De Marcos Lousa et al., 2013) and then resynthesized inside the organelle (Fulda et al., 2004). Once in the peroxisome, fatty acyl-CoAs are β-oxidized by the sequential action of three enzymes, each encoded in a small gene family in Arabidopsis (Arabidopsis thaliana): acyl-CoA oxidase (ACX1 to ACX6), multifunctional protein [MFP2, ABNORMAL INFLORESCENCE MERISTEM (AIM1)], and 3-ketoacyl-CoA thiolase (KAT1, KAT2/PED1, and KAT5). Thiolase generates acetyl-CoA and a chain-shortened fatty acyl-CoA, which can undergo additional β-oxidation rounds. Seedling acetyl-CoA is further metabolized by peroxisomal glyoxylate cycle enzymes into four-carbon dicarboxylic acids, which ultimately can be converted to Suc used for growth (for review, see Graham, 2008). Peroxisomes are the sole site of fatty acid β-oxidation in plants (for review, see Graham, 2008), and mutations in the genes encoding PXA1, various β-oxidation enzymes, or glyoxylate cycle enzymes confer seedling growth defects that are partially alleviated by providing a fixed carbon source such as Suc (for review, see Bartel et al., 2014). MFP2 and AIM1 require NAD⁺, which is imported by the PXN peroxisomal NAD⁺ carrier and recycled by peroxisomal NAD⁺-malate dehydrogenase (PMDH) and monodehydroascorbate reductase (MDAR4). Consequently, pmdh, mdar4, and pxn mutants also display β-oxidation defects (Eastmond, 2007; Pracharoenwattana et al., 2007; Bernhardt et al., 2012; Rinaldi et al., 2016; van Roermund et al., 2016).

Figure 1.

Plant peroxisome functions. Peroxisomes house a variety of catabolic and biosynthetic reactions (Reumann and Bartel, 2016), several of which generate H₂O₂ and other ROS (orange). β-oxidation (red) is used to catabolize fatty acids (purple) and in the synthesis of several hormones (blue). Peroxisomal ROS can be inactivated by catalase and other enzymes within the peroxisome or can exit the peroxisomes to serve signaling roles. OPC8, 3-oxo-2-(2′-pentenyl)-cyclopentane-1-octanoic acid; OPDA, 12-oxo-phytodienoic acid.

Oil bodies, also known as lipid droplets, compartmentalize neutral lipids and allow storage of large quantities of triacylglycerol in oilseeds (for review, see Pyc et al., 2017). During germination and early seedling growth, peroxisomes associate with oil bodies (Chapman and Trelease, 1991; Hayashi et al., 2001) to utilize seed energy reserves (for review, see Graham, 2008). Oil body triacylglycerol is hydrolyzed to free fatty acids by the lipase SUCROSE DEPENDENT1 (SDP1) (Eastmond, 2006), which is delivered from peroxisomes to oil bodies with the assistance of the Golgi retromer complex (Thazar-Poulot et al., 2015). The fatty acyl-CoA transporter PXA1 also promotes peroxisome-oil body interactions (Cui et al., 2016), possibly facilitating peroxisomal invaginations during oil body consumption.

Seedlings retain oil bodies when β-oxidation is impaired. Various mutants defective in β-oxidation, the glyoxylate cycle, or NADH reduction display peroxisomes clustered near oil bodies after oil bodies are depleted in wild type (Germain et al., 2001; Eastmond, 2007; Pracharoenwattana et al., 2007; Cassin-Ross and Hu, 2014; Rinaldi et al., 2016). Moreover, diphenyl methylphosphonate confers triacylglycerol retention (Brown et al., 2013) similar to sdp1 and pxa1 mutants (Eastmond, 2006; Slocombe et al., 2009; Kelly et al., 2013). The mechanism through which peroxisomal enzyme dysfunction feeds back to prevents triacylglycerol mobilization is unknown.

In addition to fueling germination, oil bodies and peroxisomes collaborate to provide energy in leaves. Arabidopsis CGI-58, a homolog of a mammalian lipase activator (comparative gene identification-58), promotes PXA1 function in leaves but not in germinating seeds (Park et al., 2013), and triacylglycerol accumulates in leaf oil bodies in cgi-58 mutants (James et al., 2010). Moreover, β-oxidation of triacylglycerol from stomatal oil bodies contributes to the ATP production necessary for stomatal opening upon transfer from dark to light (McLachlan et al., 2016). When β-oxidation is slowed, as in pxa1, sdp1, or cgi-58 mutants, stomatal opening is impeded (McLachlan et al., 2016). Because environmental stimuli, such as light and temperature, regulate stomatal aperture to control water and gas exchange (for review, see Hetherington and Woodward, 2003), β-oxidation-mediated stomatal opening hints that peroxisomes have a role in responding to environment cues.

β-oxidation is also important during embryogenesis. Arabidopsis pxa1 mutants and RNA interference (RNAi) lines targeting barley PXA1 homologs have small seeds (Mendiondo et al., 2014), and Arabidopsis double mutants defective in both multifunctional enzymes or several acyl-CoA oxidase isozymes die during embryogenesis (Rylott et al., 2003; Fulda et al., 2004; Khan et al., 2012). It is not known if acetyl-CoA or a different β-oxidation product (Fig. 1) is needed for embryogenesis or if β-oxidation promotes embryogenesis by preventing buildup of toxic compounds, such as free fatty acids.

The interplay between peroxisomal β-oxidation roles in providing usable fixed carbon and removing toxic free fatty acids is apparent in the multifaceted relationship between peroxisomal β-oxidation and abiotic stress survival. The transcripts encoding thiolase and ACX4 accumulate during carbon-starvation (Charlton et al., 2005; Contento and Bassham, 2010), implying that fatty acid β-oxidation increases upon nutrient deprivation. Indeed, several β-oxidation mutants, including pxa1 and mutants with reduced thiolase activity, die prematurely in extended darkness (Dong et al., 2009; Kunz et al., 2009). Both pxa1 and sdp1 mutants catabolize triacylglycerol inefficiently in extended darkness, but sdp1 lowers free fatty acid accumulation and ameliorates plant death in dark-treated pxa1 mutants (Fan et al., 2017), implying that free fatty acid toxicity rather than carbon starvation contributes to pxa1 death. Intriguingly, sdp1 survives extended darkness better than wild type (Fan et al., 2017), hinting that triacylglycerol can protect cells from darkness-induced damage.

Reactive Oxygen and Nitrogen Species: Not Just for Cell Death

The oxidative reactions harbored in peroxisomes generate hydrogen peroxide (H₂O₂) and other reactive oxygen species (ROS). In addition to the ACX β-oxidation enzymes, glycolate oxidases (GOX1 to GOX5) acting in photorespiration and xanthine oxidase acting in uric acid production contribute substantial peroxisomal H₂O₂ and superoxide radicals (for review, see Del Río and López-Huertas, 2016). Peroxisomes counter this ROS accumulation using catalase and ascorbate peroxidase pathways, which decompose H₂O₂ into water and molecular oxygen. Catalase (cat) mutants display elevated H₂O₂ and associated transcriptional changes, diminished growth, increased cell death (Queval et al., 2007), and sensitivity to carbon-starvation (Contento and Bassham, 2010). Suppressor screens for increased photosystem II efficiency in a cat2 mutant recovered a short-root mutant with decreased photorespiration flux (Waszczak et al., 2016) and a gox1 mutant (Kerchev et al., 2016), confirming photorespiratory GOX as a major H₂O₂ contributor.

Catalase has a particularly close relationship with the H₂O₂-generating ACX enzymes. For example, CAT3 and ACX4 activity and transcripts are both elevated by carbon-starvation (Contento and Bassham, 2010). Moreover, cat2 mutants display decreased ACX activity (Liu et al., 2017; Yuan et al., 2017), and overexpressing ACX3 rescues cat2 seedling growth defects (Liu et al., 2017). The finding that ACX activity is limiting for cat2 growth implies that ACX enzymes suffer damage when catalase is dysfunctional. Indeed, CAT2 physically interacts with and increases activity of ACX3 and ACX4 in vitro (Yuan et al., 2017); presumably catalase-ACX proximity allows rapid inactivation of ACX-generated H₂O₂.

In addition to ROS, peroxisomes generate reactive nitrogen species (RNS) after application of stressors such as salt or heavy metals (for review, see Corpas et al., 2017). This accumulation suggests that like ROS, RNS could function in stress signaling. The peroxisomal NADP-isocitrate dehydrogenase (pICDH) regenerates NADPH, which is used by the peroxisomal ascorbate-glutathione H₂O₂-inactivating system. picdh mutants fail to open stomates upon transfer to light unless treated with H₂O₂- or NO-scavenging chemicals (Leterrier et al., 2016), highlighting a role for peroxisomes in RNS-mediated environmental responses and providing a second example of peroxisomes influencing stomatal opening. Additionally, several peroxisomal enzymes are nitrosylated, including catalase, GOX, and PMDH (for review, see Corpas et al., 2017), which could represent RNS-mediated regulation. It will be interesting to learn the biological roles of RNS and whether RNS signals are antagonistic or agonistic to ROS signaling.

β-Oxidation: Not Just for Fatty Acids

Beyond fatty acid β-oxidation, peroxisomal enzymes β-oxidize precursors of the hormones auxin, jasmonic acid (JA), and salicylic acid (SA; Fig. 1). Indole-3-butyric acid (IBA), one of several auxin precursors in plants (for review, see Korasick et al., 2013), is converted in peroxisomes to the active auxin indole-3-acetic acid (IAA; Zolman et al., 2000; Strader et al., 2010; reviewed in Strader and Bartel, 2011). IBA-derived auxin is important during seedling development, when it influences lateral rooting (Zolman et al., 2001; De Rybel et al., 2012), cotyledon and root hair expansion, and apical hook formation (Strader and Bartel, 2009; Strader et al., 2010, 2011).

The JA precursor 12-oxo-phytodienoic acid undergoes reduction and two β-oxidation rounds in peroxisomes to yield JA (Fig. 1), which functions in reproductive development and during wound and defense responses (for review, see Wasternack and Hause, 2013). For example, a maize peroxisomal JA-modifying enzyme controls sex determination (Hayward et al., 2016). Moreover, wounding increases ACX1 and PED1 transcript levels (Cruz Castillo et al., 2004) and Arabidopsis acx1, aim1, and ped1 mutants fail to produce JA after wounding (Cruz Castillo et al., 2004; Delker et al., 2007).

Like auxin, the defense hormone SA has multiple biosynthetic pathways. SA can be produced in chloroplasts (for review, see Dempsey and Klessig, 2017) or after peroxisomal β-oxidation of transcinnamic acid to the SA precursor benzoic acid (Fig. 1). Illuminated by the pioneering work on benzoic acid biosynthesis in petunia (van Moerkercke et al., 2009; Klempien et al., 2012; Qualley et al., 2012), this pathway was elucidated in Arabidopsis. Cytosolic transcinnamic acid, presumably as the CoA ester, is imported by PXA1 (Bussell et al., 2014). Inside the peroxisome, cinnamoyl-CoA is resynthesized (Lee et al., 2012) and β-oxidized to benzoyl-CoA (Bussell et al., 2014), which is presumably hydrolyzed to benzoic acid and exported to the cytosol, where benzoic acid is converted to SA (Yalpani et al., 1993). A rice (Oryza sativa) aim1 mutant displays elevated redox gene expression, small root meristems, and short roots that are rescued by treatment with SA or ROS, but not JA or auxin, suggesting that SA promotes rice root growth via ROS production and blocking redox gene expression (Xu et al., 2017). These findings illustrate the varied means by which β-oxidation contributes to ROS and illuminate a ROS signaling role in plants.

In addition to hormone production, the peroxisome is a site of hormone cross talk. For example, SA, which is induced in response to biotrophic pathogens, directly inhibits catalase activity (Yuan et al., 2017). This catalase inhibition reduces JA production via the consequent reduction in ACX activity and reduces IAA production through H₂O₂-mediated modification of a key IAA biosynthetic enzyme (Yuan et al., 2017). Thus, SA exploits peroxisomal pathways to mediate appropriate responses to biotrophic pathogens by down-regulating the hormones (JA and IAA) that promote responses to necrotrophic pathogens.

Photorespiration: Not Just for Chloroplasts

In addition to the core processes of β-oxidation and ROS detoxification, plant peroxisomes house diverse specialized functions (for review, see Reumann and Bartel, 2016) that may change during development (Titus and Becker, 1985; Nishimura et al., 1986; Sautter, 1986; Lingard et al., 2009) or in response to environmental cues (for review, see Goto-Yamada et al., 2015). For example, plant peroxisomes sequester enzymes acting in photorespiration, which is important when ribulose-1,5-bisphosphate carboxylase/oxygenase fixes O₂ instead of CO₂. As a result, high CO₂ levels improve growth of photorespiratory-deficient plants (for review, see Timm and Bauwe, 2013). During photorespiration, peroxisomal and mitochondrial enzymes collaborate to convert glycolate from the chloroplast to glycerate to be returned to the chloroplast for the Calvin cycle. After entering the peroxisome, glycolate is oxidized by GOX to yield glyoxylate and H₂O₂ (for review, see Bauwe et al., 2010). As seedlings mature, photorespiration increases and the glyoxylate cycle diminishes, and glyoxylate is transaminated to Gly, which is converted to Ser in the mitochondrion. Ser returns to the peroxisome and is converted to glycerate by Ser:glyoxylate aminotransferase and hydroxypyruvate reductase (HPR). HPR depends on the NADH provided by PMDH (for review, see Bauwe et al., 2010).

The impaired growth of catalase mutants is ameliorated by high CO₂ (Queval et al., 2007), again implicating photorespiratory GOX as a primary H₂O₂ source in leaf peroxisomes. Knocking down both GOX1 and GOX2 confers growth defects in ambient air accompanied by decreased photosynthetic electron transfer and carbon assimilation, glycolate accumulation, and early senescence (Dellero et al., 2016). Moreover, hpr1 mutants display not only decreased photosynthetic efficiency, but also drought sensitivity (Li and Hu, 2015), linking peroxisomal photorespiration roles to drought tolerance.

PEROXISOME GENESIS

Membrane Protein Insertion and Budding from the ER

Peroxisomes are assembled and maintained by peroxin (PEX) proteins. The early acting peroxins PEX3, PEX16, and PEX19 (Fig. 2) help insert peroxisomal membrane proteins (PMPs) directly into the peroxisomal membrane (group II PMPs) or into a peroxisome-destined region of the ER membrane (group I PMPs; for review, see Hu et al., 2012). PEX16 recruits the PEX3 membrane protein to the ER in mammals (Kim et al., 2006). Neurospora PEX3 enlists PEX19, a farnesylated cytosolic protein, to chaperone nascent PMPs to PEX3 for membrane insertion (Chen et al., 2014). Yeast PMPs bind PEX19 via a membrane peroxisome-targeting signal, a hydrophobic motif near the transmembrane domain (Rottensteiner et al., 2004); similar sequences are found in plant PMPs (Nyathi et al., 2012).

Figure 2.

Peroxisome dynamics. Peroxisome biogenesis and division are coordinated by peroxins (numbered ovals) that coordinate peroxisomal membrane protein insertion into the ER or the peroxisomal membrane. After preperoxisomes bud from the ER, peroxisomes mature through import of matrix proteins. Peroxisomes can be degraded by pexophagy, a type of specialized autophagy. Dynamic peroxisome extensions (peroxules) assist peroxisome interactions with other organelles and can be associated with peroxisome division. PEX11 promotes peroxisome division together with a group of proteins (PMD1, FIS1, DRP) that also act in division of mitochondria or chloroplasts. PMP, peroxisomal membrane protein.

In addition to PMP insertion, yeast PEX3 and PEX19 are implicated in budding of ER-derived preperoxisomal vesicles (van der Zand et al., 2010) carrying distinct PMP assortments (Agrawal et al., 2016). Moreover, mammalian PEX3 can be inserted into the mitochondrial outer membrane, and mitochondrion-derived preperoxisomal vesicles can fuse with PEX19-containing ER-derived preperoxisomal vesicles to form import-competent peroxisomes (Sugiura et al., 2017).

Much remains to be discovered about peroxisome biogenesis in plants. Like in mammals, Arabidopsis PEX16 is delivered to the peroxisome via the ER, where it recruits other PMPs (Hua et al., 2015). Arabidopsis PEX16 RNAi lines display large peroxisomes and slightly impaired β-oxidation (Nito et al., 2007), and an insertional pex16 allele displays severe embryonic defects (Lin et al., 1999). Arabidopsis has two isoforms of PEX3 and PEX19. Single pex19 insertional alleles lack obvious defects whereas a pex19a pex19b double mutant is embryo-lethal, indicating functional redundancy (McDonnell et al., 2016). PEX3 or PEX19 RNAi lines display large peroxisomes but wild-type β-oxidation (Nito et al., 2007). The composition, organellar origins, and fusion mechanisms of plant preperoxisomal vesicles remain to be elucidated.

Peroxisome Division and Proliferation

In addition to budding from the ER, peroxisomes can divide by fission (Fig. 2). Plant peroxisomes proliferate during cell division (Lingard et al., 2008) and in response to salinity (Mitsuya et al., 2010; Fahy et al., 2017), light (Desai and Hu, 2008), and cadmium treatments (Rodríguez-Serrano et al., 2016). Division involves the PMP PEX11, which has five isoforms (a to e) in Arabidopsis (Lingard and Trelease, 2006). Although decreasing PEX11 expression via RNAi does not notably impact β-oxidation or matrix protein import, Arabidopsis pex11 RNAi lines (Nito et al., 2007; Orth et al., 2007) and moss pex11 mutants (Kamisugi et al., 2016) exhibit enlarged peroxisomes, suggesting a conserved division role. Additionally, PEX11a is implicated in forming peroxisomal membrane extensions called “peroxules” (Rodríguez-Serrano et al., 2016). Peroxule formation is induced by ROS (Sinclair et al., 2009; Rodríguez-Serrano et al., 2016) and may promote the peroxisomal elongation that precedes division (Fig. 2). Furthermore, loss of PEX11a decreases catalase and superoxide dismutase gene expression, linking ROS signaling and peroxisomal division (Rodríguez-Serrano et al., 2016).

After elongation, several proteins collaborate to divide the peroxisome (Fig. 2). The Arabidopsis paralogs of yeast FISSION1 (Kemper et al., 2008), FIS1A and FIS1B, are tail-anchored membrane proteins acting in both mitochondrial and peroxisomal fission (for review, see Hu et al., 2012). Knocking down FIS1A and FIS1B decreases peroxisome numbers in protoplasts (Lingard et al., 2008), and insertional fis1a alleles display larger and fewer peroxisomes (Zhang and Hu, 2009) and mitochondria (Scott et al., 2006) than wild type.

The dynamin-related proteins, DRP3A, DRP3B, and DRP5B, are GTPases that, like FIS1, are required for division of multiple organelles. DRP3 functions in peroxisomal and mitochondrial fission whereas DRP5B supports fission of peroxisomes and chloroplasts (for review, see Hu et al., 2012). Arabidopsis drp3a and drp3b mutants both display larger and fewer mitochondria, but only drp3a displays larger and fewer peroxisomes (Mano et al., 2004; Fujimoto et al., 2009; Zhang and Hu, 2009) coupled with slight β-oxidation defects (Mano et al., 2004). Overexpressing DRP3B, but not DRP3A, causes peroxisome elongation (Fujimoto et al., 2009), suggesting that DRP3B promotes elongation whereas DRP3A functions in constriction and scission. Null drp5b alleles display larger and clustered peroxisomes, slight β-oxidation defects, and growth defects rescued by high CO₂ (Zhang and Hu, 2010).

PEX11s might recruit other fission machinery to the peroxisome once elongation has commenced. All five Arabidopsis PEX11 isoforms can bind FIS1A (Lingard et al., 2008), and moss PEX11 and FIS1A interact at the peroxisomal membrane (Kamisugi et al., 2016). Moreover, Arabidopsis DRP5B binds PEX11s as well as FIS1A, DRP3A, and DRP3B (Zhang and Hu, 2010).

The plant-specific PEROXISOMAL AND MITOCHONDRIAL DIVISION1 (PMD1) is a tail-anchored membrane protein that acts independently of PEX11s, FIS1s, and DRPs to promote peroxisome and mitochondrial division (Aung and Hu, 2011). pmd1 mutants display elongated mitochondria and larger and fewer peroxisomes than wild type (Aung and Hu, 2011). Like PEX11 (Mitsuya et al., 2010), PMD1 promotes peroxisome proliferation in response to salt (Frick and Strader, 2017), although this proliferation does not seem to impact salt tolerance (Mitsuya et al., 2010; Frick and Strader, 2017). Interestingly, salt-induced proliferation also requires MAP Kinase 17 (Frick and Strader, 2017), implying a role for phosphorylation in peroxisome proliferation.

MATRIX PROTEIN IMPORT: CYCLING RECEPTORS

Cargo Selection by PTS1 and PTS2 Receptors

Matrix protein import (Fig. 3) replenishes peroxisomal contents and converts preperoxisomes to mature peroxisomes (Fig. 2). Two types of peroxisome targeting signals (PTS) specify matrix protein localization. Most matrix proteins carry a PTS1, a C-terminal SKL or similar tripeptide (Reumann, 2004; Lingner et al., 2011). Fewer proteins carry the PTS2 nonapeptide, often R[L/I]X₅HL in plants, near the N terminus (Reumann, 2004). After delivery, the PTS1 region is retained whereas the approximately 30-amino acid N-terminal region of plant PTS2 proteins is cleaved by the protease DEG15 (Fig. 3; Helm et al., 2007; Schuhmann et al., 2008). Although plants, yeast, and mammals use both PTS1 and PTS2 systems, nematodes and fruit flies lack PTS2 proteins (Gurvitz et al., 2000; Motley et al., 2000; Faust et al., 2012).

Figure 3.

Matrix protein import and receptor recycling. Matrix proteins harboring peroxisome-targeting signals are synthesized in the cytosol, where they are recognized by the PEX5 (PTS1 proteins) or PEX7 (PTS2 proteins) receptors. Receptor-cargo complexes dock with PEX13-PEX14, which allows cargo release into the matrix. Membrane-associated PEX5 is ubiquitinated near the N-terminus by enzymes in the RING complex assisted by the PEX4 ubiquitin-conjugating enzyme. Monoubiquitinated or diubiquitinated PEX5 is recycled via removal from the membrane by the PEX1-PEX6 ATPase complex whereas PEX5 polyubiquitination can lead to PEX5 proteasomal degradation or may trigger pexophagy. PTS2 proteins are processed in the matrix by the DEG15 protease. C, C-terminus; N, N-terminus; Ub, ubiquitin.

Several algorithms predict plant PTS1 proteins, including PredPlantPTS1 (Reumann et al., 2012) and PPero (Wang et al., 2017). Bioinformatic and proteomic approaches have identified many potential peroxisomal proteins in plants (for review, see Reumann, 2011). These analyses have uncovered noncanonical PTS1 signals and revealed the importance of residues upstream of the PTS1 for targeting (Chowdhary et al., 2012). As not all predicted targeting signals confer peroxisomal localization (Ching et al., 2012), fusions of fluorescent reporters to candidate matrix proteins can be used to visualize localization in transgenic plants (Mano et al., 1999; Cassin-Ross and Hu, 2014; Wu et al., 2016) or following transient transfection of tobacco leaves (Reumann et al., 2009; Quan et al., 2013), cell culture (Mano et al., 1999; Carrie et al., 2007), or onion epidermal cells (Chowdhary et al., 2012; Skoulding et al., 2015).

PTS1 proteins are recognized by PEX5 (van der Leij et al., 1993; Zolman et al., 2000), and PTS2 proteins are recognized by PEX7 (Fig. 3; Marzioch et al., 1994; Braverman et al., 1997; Woodward and Bartel, 2005). Yeast PEX7 contains six WD40 domains forming a seven-bladed propeller that binds the PTS2 peptide on one face of PEX7 (Pan et al., 2013). The C-terminal region of PEX5 contains two clusters of tetratricopeptide repeats that bind the PTS1 (Gatto et al., 2000; Hagen et al., 2015). The strength of in vitro binding of PTS1 variants to PEX5 correlates with in vivo targeting efficiency in higher plants (Skoulding et al., 2015). Peroxisomal constituents also may affect import. For example, nitric oxide donors and a calmodulin antagonist impair Arabidopsis PTS1 import, implicating nitric oxide and calcium as import regulators (Corpas and Barroso, 2017).

Interestingly, peroxisomes can import folded and oligomeric proteins (McNew and Goodman, 1994; Lee et al., 1997), which allows some endogenous proteins lacking a PTS to “piggyback” into peroxisomes (Kataya et al., 2015). However, the import machinery prefers monomeric proteins (Freitas et al., 2015), and PEX5 binding to catalase (Freitas et al., 2011), acyl-CoA oxidase1, and urate oxidase (Freitas et al., 2015) prevents oligomerization of these cargo proteins.

As in mammals (Braverman et al., 1998; Otera et al., 1998), PEX7-PEX5 interactions allow PTS2 protein delivery in plants (Hayashi et al., 2005; Woodward and Bartel, 2005). In humans, alternative splicing produces two PEX5 forms: a short form competent for PTS1 import and a long form facilitating both PTS1 and PTS2 import (Dodt et al., 1995; Braverman et al., 1998). Although only one Arabidopsis PEX5 splice form is reported, rice contains alternative forms, and only the long form binds PEX7 (Lee et al., 2006). PEX5 and PEX7 may interact via several regions. The PEX5 N-terminal region (1 to 230 amino acids) binds PEX7 in yeast two-hybrid assays (Nito et al., 2002), and an Arabidopsis pex5 variant lacking residues 314 to 334 fails to bind PEX7 in pull-down assays (Lanyon-Hogg et al., 2014). The pex5-10 mutant and PEX5 RNAi lines display β-oxidation defects and impaired import of both PTS1 and PTS2 proteins (Hayashi et al., 2005; Zolman et al., 2005; Khan and Zolman, 2010). Expressing an N-terminal PEX5 domain in pex5-10 restores PTS2 processing, showing that the PEX5 N-terminal domain promotes PEX7 function in vivo (Khan and Zolman, 2010). Moreover, a special Arabidopsis pex5-1 (S318L) missense mutation confers inefficient β-oxidation and PTS2 import but normal PTS1 import (Zolman et al., 2000; Woodward and Bartel, 2005).

Arabidopsis pex7 mutants display β-oxidation and PTS2 import defects (Hayashi et al., 2005; Woodward and Bartel, 2005; Ramón and Bartel, 2010). Surprisingly, several pex7 mutations also impair PTS1 import and lower PEX5 levels (Ramón and Bartel, 2010), revealing that PEX7 promotes PEX5 stability.

In addition to targeting, PEX5-PEX7 interactions may influence cargo unloading. Structural studies of Saccharomyces cerevisiae peroxins reveal the PTS2 peptide sandwiched between PEX7 and its coreceptor PEX21 (Pan et al., 2013), which in yeast functions like plant PEX5 to bring PEX7 to the organelle. Perhaps PEX5 conformational changes during membrane insertion or PTS1 cargo unloading reconfigure PEX5-PTS2 cargo-PEX7 interactions to promote PTS2 cargo unloading.

Docking Receptor-Cargo Complexes at the Peroxisome

The receptor-cargo complex docks with PEX13 and PEX14 on the peroxisomal membrane. In yeast, PEX5 and PEX14 form a dynamic translocation pore with a cargo-dependent diameter (Meinecke et al., 2010). In plants, the PEX14 N-terminal region binds PEX5 WXXXF/Y domains (Nito et al., 2002), in vitro label transfer assays implicate PEX14 as the first peroxisomal contact of PEX5 during import (Bhogal et al., 2016), and pex14 mutants display impaired β-oxidation and matrix protein import (Hayashi et al., 2000; Monroe-Augustus et al., 2011; Burkhart et al., 2013). However, Arabidopsis pex14 null alleles are viable (Monroe-Augustus et al., 2011), whereas pex13 null alleles confer lethality (Boisson-Dernier et al., 2008), hinting that some yeast PEX14 roles might be provided by PEX13 in plants. PEX13 dysfunction results in expected physiological defects; a pex13 RNAi line and two missense pex13 mutants, aberrant peroxisome morphology 2 (apm2) and pex13-4, display β-oxidation and matrix protein import defects (Mano et al., 2006; Nito et al., 2007; Woodward et al., 2014). Moreover, the pex13-4 mutation lowers PEX5 membrane association, and PEX5 overexpression ameliorates a subset of pex13-4 defects (Woodward et al., 2014), implying that the pex13-4 matrix protein import defects are due to impaired PEX5 docking.

PEX13 binds PEX14 in yeast (Pires et al., 2003) and mammals (Fransen et al., 1998), but this interaction has not been reported in plants. Yeast PEX13 interacts with PEX14 via a C-terminal Src homology 3 (SH3) domain and an intraperoxisomal sequence; this interaction is essential for matrix protein import (Schell-Steven et al., 2005). PEX13 also binds PEX5 and PEX7 in yeast (Douangamath et al., 2002; Stein et al., 2002; Pires et al., 2003) and mammals (Otera et al., 2002). Although Arabidopsis PEX13 does bind to PEX7 (Mano et al., 2006), Arabidopsis PEX13 lacks a recognizable SH3 domain (Boisson-Dernier et al., 2008), and PEX5-PEX13 interactions have not been reported in plants (Mano et al., 2006). It remains to be determined if these apparent receptor docking differences reflect functional diversity or technical challenges.

In addition to recruiting cargo-receptor complexes to peroxisomes, docking complex-receptor interactions may promote cargo unloading. In Pichia pastoris, PTS1 cargo binding enhances PEX5-PEX14 interaction but weakens PEX5-PEX13 interaction (Urquhart et al., 2000), suggesting that PEX14 initiates docking, and PEX13 promotes PTS1 cargo release. Moreover, the N-terminal region of Arabidopsis PEX14 is sufficient to isolate PEX5 and PEX7 but not PTS2 cargo (Lanyon-Hogg et al., 2014), suggesting that PEX14 binding might promote PTS2 cargo unloading.

Roles for Ubiquitination in Receptor Recycling and Peroxin Degradation

After cargo delivery, ubiquitination promotes the recycling of cargo receptors from the peroxisomal membrane back to the cytosol (Fig. 3). During ubiquitination, ubiquitin-conjugating enzymes (UBCs) assist ubiquitin-protein ligases in covalently attaching ubiquitin to substrate proteins. S. cerevisiae PEX5 monoubiquitination by the peroxisome-tethered UBC PEX4 and the peroxisomal ubiquitin-protein ligase PEX12 (Platta et al., 2009) allows a peroxisome-tethered ATPase complex to recycle PEX5 to the cytosol for further rounds of cargo recruitment (for review, see Grimm et al., 2012). In contrast, PEX5 polyubiquitination by the cytosolic UBC4 acting with the peroxisomal ubiquitin-protein ligase PEX2 targets PEX5 for proteasomal degradation (Platta et al., 2009). The role of the third RING peroxin, PEX10, is controversial. Mammalian PEX10 is essential (Okumoto et al., 2014), but yeast PEX10 only enhances PEX5 ubiquitination (Platta et al., 2009; El Magraoui et al., 2012).

Although PEX5 ubiquitination has not been directly demonstrated in plants, mutants defective in the peroxisome-associated ubiquitination machinery reveal roles in plant growth, peroxisomal import, and PEX5 retrotranslocation. The pex4-1 missense mutant and pex4 RNAi lines show impaired β-oxidation and matrix protein import (Zolman et al., 2005; Nito et al., 2007). PEX5 accumulates (Kao et al., 2016) and is excessively membrane-associated (Ratzel et al., 2011; Kao and Bartel, 2015) in pex4-1, indicating that PEX4 promotes both PEX5 degradation and PEX5 retrotranslocation. Moreover, overexpressing PEX5 exacerbates pex4-1 defects (Kao and Bartel, 2015), suggesting that PEX5 retention in the peroxisomal membrane is detrimental. Interestingly, a T-DNA insertion upstream of the PEX13 start codon (pex13-1) that lowers PEX13 transcripts alleviates pex4-1 growth defects (Ratzel et al., 2011). This suppression implies that decreasing receptor docking lessens the detrimental effects of PEX5 retention. Similarly, growth at elevated temperature lowers PEX5 levels and alleviates the peroxisomal defects in pex4 mutants (Kao and Bartel, 2015).

PEX22 tethers PEX4 to the peroxisome (Fig. 3). Arabidopsis PEX22 was identified via its PEX4-binding ability and can function in yeast when expressed together with Arabidopsis PEX4 (Zolman et al., 2005). Yeast PEX22 enhances PEX4 enzymatic activity (El Magraoui et al., 2014), and a T-DNA insertion upstream of the Arabidopsis PEX22 start codon exacerbates the peroxisomal defects of pex4-1 (Zolman et al., 2005).

The Arabidopsis PEX2, PEX10, and PEX12 RING peroxins all display in vitro ubiquitin-protein ligase activity (Kaur et al., 2013) and are essential for embryogenesis (Hu et al., 2002; Schumann et al., 2003; Sparkes et al., 2003; Fan et al., 2005; Prestele et al., 2010). Expressing truncated RING peroxins without the C-terminal catalytic zinc-binding RING domains (ΔZn) in wild type confers dominant-negative matrix protein import defects for PEX2-ΔZn and photorespiration defects attributed to decreased peroxisome-chloroplast interactions for PEX10-ΔZn (Prestele et al., 2010). RNAi lines targeting RING peroxin genes (Nito et al., 2007) and several viable RING peroxin mutants (Mano et al., 2006; Burkhart et al., 2014; Kao et al., 2016) show typical peroxisomal defects, including impaired β-oxidation and matrix protein import. Moreover, PTS1 and PTS2 receptor levels are increased in RING peroxin mutants (Kao et al., 2016), and PEX5 is excessively membrane-associated in a pex12 mutant (Mano et al., 2006), suggesting that the RING peroxins facilitate PEX5 and PEX7 retrotranslocation.

Both Arabidopsis pex12 missense mutants are partial loss-of-function alleles with Lys substitutions at adjacent amino acid residues (R170K in apm4 and E171K in pex12-1) in a relatively nonconserved 49 amino acid region lacking Lys residues (Mano et al., 2006; Kao et al., 2016). Surprisingly, reducing PEX4 function ameliorates rather than exacerbates pex12-1 peroxisomal defects (Kao et al., 2016). This suppression suggests that the pex12-1 ectopic Lys residue might provide an attachment site for PEX4-assisted ubiquitination, triggering degradation of the pex12 protein.

The RING peroxins form a complex, and each component contributes to complex stability in yeast (Hazra et al., 2002; Agne et al., 2003; Okumoto et al., 2014). Similarly, Arabidopsis pex2-1, pex10-2, and pex12-1 mutants all display decreased PEX10 levels (Kao et al., 2016). Along with physiological restoration, pex4 mutants restore PEX10 levels in pex12-1 (Kao et al., 2016). Thus, both PEX10 and PEX12 could be substrates, along with PEX5, of the peroxisomal ubiquitination machinery.

The RING peroxins may not be the only peroxisome-associated ubiquitin-protein ligases. The suppressor of plastid protein import locus 1 (SP1) is a RING-type ubiquitin-protein ligase localizing on chloroplasts, where it promotes degradation of several outer envelope translocon components (Ling et al., 2012) and modulates abiotic stress tolerance (Ling and Jarvis, 2015). A recent report suggests that SP1 also can localize to peroxisomes and interact with the docking peroxins, where it promotes PEX13 ubiquitination and degradation (Pan et al., 2016). Loss of SP1 increases β-oxidation in wild type and improves peroxisome function in pex13-1 and pex14-2 mutants (Pan et al., 2016). Interestingly, sp1 mutants exacerbate pex4-1 defects (Pan et al., 2016), consistent with the hypothesis that excessive docking capacity is detrimental when PEX5 recycling is impaired (Ratzel et al., 2011). However, SP1 peroxisomal localization may depend on overexpression, and PEX13 and PEX14 levels do not consistently vary with SP1 accumulation in seedlings (Ling et al., 2017), highlighting the possibility that peroxisome-related sp1 phenotypes may be indirect effects of altered chloroplast function.

Like SP1, PEX2 may impact both chloroplasts and peroxisomes. A pex2 missense allele (ted3) suppresses the photomorphogenic defects of the de-etiolated1 (det1) mutant (Hu et al., 2002), and expressing a GFP-fused PEX2 RING domain slightly ameliorates det1 growth defects (Desai et al., 2014). Many metabolic pathways are shared among organelles. For example, photorespiration requires enzymes acting in peroxisomes, chloroplasts, and mitochondria, suggesting that additional shared regulatory machinery awaits discovery.

Recycling of the PTS2 receptor PEX7 is not well understood. In mammals, PEX7 export requires PEX5 export (Rodrigues et al., 2014), and dysfunctional PEX7 is ubiquitinated and degraded (Miyauchi-Nanri et al., 2014). Disrupting PEX5 recycling increases PEX7 levels in P. pastoris (Hagstrom et al., 2014) and Arabidopsis (Kao et al., 2016), suggesting coordinated degradation. In addition, two Arabidopsis Rab GTPases bind GFP-PEX7 on the peroxisomal membrane and promote proteasomal degradation of membrane-associated PEX7 (Cui et al., 2013). Whether these Rab GTPases impact PEX5 recycling or the peroxisomal ubiquitination machinery is unknown.

ATP-Dependent Receptor Retrotranslocation

Monoubiquitinated PEX5 is returned to the cytosol by a peroxisome-tethered ATPase complex (Fig. 3). PEX1 and PEX6 are members of the ATPases associated with diverse cellular activities family and function in yeast as a trimer of PEX1-PEX6 dimers (Blok et al., 2015; Ciniawsky et al., 2015; Gardner et al., 2015). The PEX1-PEX6 heterohexamer is tethered to the peroxisome by a tail-anchored protein known as PEX15 in yeast (Elgersma et al., 1997), PEX26 in mammals (Matsumoto et al., 2003), and APEM9/DAYU/PEX26 in plants (Goto et al., 2011; Li et al., 2014; Gonzalez et al., 2017). PEX26 binds PEX1-PEX6 via PEX6 (Birschmann et al., 2003; Matsumoto et al., 2003; Goto et al., 2011). Unlike PEX22 enhancement of PEX4 activity (El Magraoui et al., 2014), tether binding decreases PEX1-PEX6 ATPase activity in yeast (Gardner et al., 2015). In addition to tethering PEX1-PEX6, mammalian PEX26 interacts with the PEX14 docking peroxin (Tamura et al., 2014), hinting that PEX26 may bridge the import and export machinery.

Arabidopsis RNAi lines targeting PEX1, PEX6, or PEX26 display decreased β-oxidation and matrix protein import (Nito et al., 2007; Goto et al., 2011). Although PEX1 is the most commonly mutated gene in peroxisome biogenesis disorder patients (for review, see Braverman et al., 2016), Arabidopsis pex1 mutants were only recently reported (Rinaldi et al., 2017). pex1-3 is inviable when homozygous and displays impaired matrix protein import and enlarged peroxisomes when heterozygous (Rinaldi et al., 2017). pex1-2 displays impaired matrix protein import and low levels of both PEX1 and PEX6 (Rinaldi et al., 2017), suggesting that PEX1 normally stabilizes PEX6. Overexpressing PEX6 restores PEX1 levels and ameliorates pex1-2 peroxisomal defects (Rinaldi et al., 2017), suggesting reciprocal stabilization of PEX1 by PEX6.

Four Arabidopsis pex6 mutants have been characterized. pex6-1, pex6-3, and pex6-4 alter residues near the second ATPase domain (Zolman and Bartel, 2004; Gonzalez et al., 2017) and display decreased β-oxidation, delayed oil body utilization, impaired matrix protein import, low PEX5 levels (Zolman and Bartel, 2004; Gonzalez et al., 2017), and increased PEX5 membrane association (Ratzel et al., 2011; Gonzalez et al., 2017), implying that PEX5 is degraded when recycling is impaired (Fig. 3). The atypical pex6-2 mutant displays elevated PEX5 levels and delayed matrix protein degradation but resembles wild type in most measures of peroxisome function (Burkhart et al., 2013; Gonzalez et al., 2017).

Arabidopsis pex26 null mutants display embryo lethality (Goto et al., 2011) and pollen maturation defects (Li et al., 2014). The viable aberrant peroxisome morphology9 missense allele shows wild-type β-oxidation but impaired matrix protein import in some cells (Goto et al., 2011). The pex26-1 splice-site mutation confers β-oxidation deficiency and low PEX5 levels like typical pex6 mutants (Gonzalez et al., 2017). Mutations in PEX4 or RING peroxins restore PEX5 levels in pex26-1 (Gonzalez et al., 2017), and a pex4 mutant restores PEX5 levels in pex6-1 (Ratzel et al., 2011), suggesting that ubiquitination triggers the heightened PEX5 degradation observed in these mutants. Together, the evidence suggests that ubiquitination drives PEX5 recycling or degradation in plants as in other eukaryotes (Fig. 3), but direct demonstration of PEX5 ubiquitination in plants would bolster this conclusion.

Overexpressing PEX5 worsens the peroxisomal defects of pex1-2 (Rinaldi et al., 2017), pex4-1 (Kao and Bartel, 2015), pex6-2 (Burkhart et al., 2013), pex6-4 (Gonzalez et al., 2017), and pex26-1 (Gonzalez et al., 2017), suggesting that PEX5 impedes peroxisome function when not efficiently recycled. In contrast, overexpressing PEX5 ameliorates pex6-1 (Zolman and Bartel, 2004) and pex6-3 (Gonzalez et al., 2017) defects. These differences hint that the PEX1-PEX6 complex may retrotranslocate not only monoubiquitinated PEX5 for recycling but perhaps also polyubiquitinated substrates for proteasomal degradation (Gonzalez et al., 2017).

QUALITY CONTROL AND PEXOPHAGY

Peroxisomes house many oxidative reactions (Fig. 1), and although antioxidative enzymes can detoxify ROS, peroxisomes and their constituents are still likely to be damaged and require turnover. Eukaryotes dispose of large cytosolic components, including organelles, via autophagy (for review, see Li and Vierstra, 2012). Peroxisome turnover is mediated by selective autophagy of peroxisomes, or pexophagy (for review, see Young and Bartel, 2016).

Various organisms use different signals to recruit autophagy receptors during pexophagy (for review, see Honsho et al., 2016), complicating the search for pexophagy-specific machinery in Arabidopsis. In Hansenula polymorpha, PEX14 is the only peroxin required for pexophagy (Zutphen et al., 2008). In S. cerevisiae, PEX3 recruits a yeast-specific autophagy-related protein, ATG36, to target the organelle for degradation (Motley et al., 2012). In mammals, Neighbor of BRCA1 Gene 1 and p62 trigger pexophagy by linking the autophagy machinery to ubiquitinated proteins on the peroxisome surface (Deosaran et al., 2013); expressing a cytosol-facing ubiquitin-tagged PMP is sufficient to trigger pexophagy (Kim et al., 2008). PEX2-mediated ubiquitination of PEX5 or PMP70 triggers pexophagy during starvation (Sargent et al., 2016) and ROS increase PEX5 phosphorylation, leading to PEX5 ubiquitination and subsequent p62-mediated pexophagy (Zhang et al., 2015).

Arabidopsis pexophagy was recently demonstrated (Farmer et al., 2013; Kim et al., 2013; Shibata et al., 2013). During seedling development, peroxisome functions shift from fatty acid utilization to photorespiration (Titus and Becker, 1985; Nishimura et al., 1986; Sautter, 1986; Lingard et al., 2009). Autophagy mutants accumulate peroxisomal proteins (Shibata et al., 2013; Yoshimoto et al., 2014) and peroxisomes (Kim et al., 2013; Yoshimoto et al., 2014) during this transition, suggesting a role for pexophagy in clearing obsolete peroxisomes. Moreover, autophagy-defective mutants were recovered in a microscopy-based screen for aggregated peroxisomes (Shibata et al., 2013). H₂O₂ treatment or reducing catalase function also results in peroxisome clustering in autophagy-defective mutants (Shibata et al., 2013; Yoshimoto et al., 2014). These findings suggest that oxidatively damaged peroxisomes are degraded via autophagy.

The autophagy machinery coordinates peroxisome abundance together with the peroxisomal matrix protease LON2 (Farmer et al., 2013). The chaperone activity of LON2 normally inhibits pexophagy (Goto-Yamada et al., 2014), and as cells age, lon2 mutants develop β-oxidation defects and low peroxisomal protein levels (Lingard and Bartel, 2009) due to heightened pexophagy (Farmer et al., 2013).

Interestingly, lon2 and PEX1/pex1-3 peroxisomes appear enlarged, and preventing autophagy restores peroxisome size in both mutants (Farmer et al., 2013; Goto-Yamada et al., 2014; Rinaldi et al., 2017), suggesting that these enlarged peroxisomes are pexophagy intermediates. PEX1 dysfunction in yeast (Nuttall et al., 2014) and mammalian cells (Law et al., 2017) also triggers pexophagy. These findings imply that LON2 and/or PEX1-PEX6 clients, perhaps including ubiquitinated PEX5, promote pexophagy in plants.

Autophagy receptors often bind the ubiquitin-like protein ATG8, which decorates the growing autophagosome membrane (for review, see Li and Vierstra, 2012). Intriguingly, the Arabidopsis RING peroxin PEX10 and the ATPase PEX6 bind ATG8 in bimolecular fluorescence complementation assays (Xie et al., 2016). Moreover, Arabidopsis DSK2, a ubiquitin-binding protein that interacts with the RING domains of PEX2 and PEX12 (Kaur et al., 2013), also binds ATG8 and promotes selective autophagy of a growth-promoting transcription factor (Nolan et al., 2017). Characterizing pexophagy in pex or dsk2 mutants might assist in identifying the molecular triggers and receptors for pexophagy in plants.

FUTURE PERSPECTIVES

Although our understanding of plant peroxisome biology is expanding, much remains to be discovered (see Outstanding Questions). The enzymes catalyzing peroxisomal fatty acid metabolism, photorespiration, and ROS inactivation are identified, but how matrix protein levels are controlled, how metabolites leave the organelle, how peroxisomes function as both sources and sinks of ROS and RNS, and how peroxisome-derived ROS and RNS integrate with signals from other organelles, remain mysterious.

How peroxisome biogenesis from the ER is balanced with division of existing organelles is an open question. In addition to our limited understanding of peroxisome biogenesis from the ER, the proteins implicated in plant peroxisome division are redundantly encoded in plants and often also participate in division of mitochondria or chloroplasts, making it challenging to isolate the roles of peroxisome division in plant physiology. Moreover, although the peroxins that directly mediate peroxisome biogenesis and division are identified, the transcriptional regulation of plant PEX genes is understudied, and only a few factors involved in PEX11 expression are identified (Desai and Hu, 2008; Desai et al., 2017).

Although peroxins were initially discovered due to their roles in peroxisome biogenesis, additional functions for these proteins continue to emerge. The peroxins that mediate PEX5 ubiquitination and retrotranslocation (Fig. 3) resemble enzymes acting in ER-associated degradation (for review, see Schliebs et al., 2010), and evidence is mounting that these receptor-recycling peroxins may ubiquitinate and remove additional clients from the peroxisomal membrane (Burkhart et al., 2014; Kao et al., 2016; Gonzalez et al., 2017). Mammalian PEX3 and PEX19 function not only in PMP insertion but also in inserting the lipid droplet- and ER-associated hairpin protein, UBXD8 (Schrul and Kopito, 2016). Moreover, mammalian PEX3 and PEX13 promote autophagy of mitochondria (mitophagy) whereas PEX19 and PEX14 are necessary for general autophagy (Lee et al., 2017). The dual roles of peroxins acting in biogenesis and to attract autophagy machinery (Zutphen et al., 2008; Motley et al., 2012; Xie et al., 2016) hint at mechanisms to trigger peroxisome degradation when import becomes dysfunctional. These discoveries highlight the intimate relationships among organelles and prompt the question of whether plant peroxins are similarly promiscuous.

Given the close metabolic connections between peroxisomes and other organelles, it is not surprising that tight physical associations are observed, for example, among peroxisomes and the ER (Barton et al., 2013) and chloroplasts (Schumann et al., 2007; Oikawa et al., 2015). Peroxules can mediate interorganellar contacts, such as among peroxisomes and ER (Sinclair et al., 2009), oil bodies (Thazar-Poulot et al., 2015), mitochondria (Jaipargas et al., 2016), and chloroplasts (Gao et al., 2016). Moreover, peroxules can respond to environmental signals. For example, peroxules are induced by oxidative stress (Sinclair et al., 2009), and high light rapidly induces peroxule interactions with mitochondria (Jaipargas et al., 2016). The study of peroxule dynamics is in its infancy, and how proteins on the peroxisome and target organelle mediate these interactions awaits discovery.

Finally, much of what we know about plant peroxisome biogenesis and function comes from research using the reference plant Arabidopsis. Additional genetic investigations in other plants, including in nonoilseed crop plants (Mendiondo et al., 2014) and nonflowering plants (Kamisugi et al., 2016), are needed to understand the diverse roles and regulation of peroxisomes throughout the plant kingdom. New chemical tools to visualize (Landrum et al., 2010; Fahy et al., 2017) and disrupt (Brown et al., 2011, 2013) plant peroxisomes will likely accelerate these studies.