Stress induces peroxisome biogenesis genes

PEX1 from Arabidopsis

We cloned a PEX1 cDNA from Arabidopsis. A composite sequence (DDBJ/EMBL/GenBank accession No. AF275382) was assembled from two clones that are identical in 189 bp of overlap. The 5′-end of the cDNA sequence and ∼995 bp of genomic sequence are shown in Figure 1. There are two in-frame ATG motifs (underlined). The more probable start site is the second ATG as it possesses a strong Kozak consensus GTGATGG (Kozak, 1997). Assuming that translation starts at this ATG, the deduced amino acid sequence of Arabidopsis Pex1p comprises 1117 amino acids with a predicted molecular weight of 122 305 Da. Two ATP-binding motifs (P-loops) are predicted between amino acids 604 and 611 and 889 and 896, and an AAA family signature between 985 and 1003. A protein database search with AtPex1p identified human Pex1p as the most similar sequence in the database (E = 1 × 10^–79), closely followed by Pex1p from the yeasts Yarrowia lipolytica, H.polymorpha and P.pastoris. AtPex1p shares identity with known Pex1 proteins ranging from 27.6% with S.cerevisiae to 29.3% with P.pastoris. AtPex1p can be immunoprecipitated by antibodies raised against P.pastoris Pex1p (data not shown). The PEX1 promoter does not contain any obvious ORE or PPRE-like elements.

AtPEX1 expression in seedlings

Figure 2 shows the expression profile of AtPEX1 and AtPEX5 (Brickner et al., 1998), which encodes Pex5p, the receptor for PTS1-targeted proteins (Dodt et al., 1995). Both genes are expressed at a low level in imbibed seeds and increase rapidly during the first 2 days of post-germinative growth before falling to a lower steady-state level by day 3, which is maintained to day 8. At this time (0–3 days post-imbibition), peroxisome size and number increase considerably in the cotyledon cells of Arabidopsis (Mansfield and Briarty, 1996). Malate synthase, a unique enzyme of the glyoxylate cycle, is maximally expressed at 1–2 days post-imbibition, which coincides with the maximal rate of lipid breakdown (Mansfield and Briarty, 1996), but thereafter declines to almost undetectable levels (Eastmond et al., 2000). The up-regulation of PEX1 and PEX5 expression at days 1–2 is consistent with their involvement in peroxisome biogenesis, and their continued expression at a lower level after day 2 presumably reflects the requirement for continued peroxisome biogenesis as cells divide and organelles turn over.

PEX1 induction by hydrogen peroxide

Different stress situations yield hydrogen peroxide as a signal molecule that triggers gene expression of various cellular protectants (Levine et al., 1994; Mittler and Zilinkas, 1994; Willekens et al., 1994a; Karpinski et al., 1999; Guan et al., 2000). As proliferation of peroxisomes and induction of some peroxisomal genes in response to oxidative stress has been reported, the effect of the stress signal molecule H₂O₂ on the expression of PEX genes was investigated. Northern blots showed that AtPEX1 was induced ∼10-fold in detached leaves by H₂O₂ (Figure 3A), similar to the level of induction of glutathione S-tranferase (GST), a well-characterized H₂O₂-responsive gene (Levine et al., 1994). The same result was obtained by semi-quantitative RT–PCR (Figure 3B) with gene-specific primers against AtPEX1, GST, AtPEX5, AtPEX14 (J.Oh, E.Lopez-Huertas, W.Charlton and A.Baker, in preparation) and AtPEX10 (Schuman et al., 1999). As peroxisomes may also be important in protecting animal cells against oxidative stress, the effect of H₂O₂ treatment on PEX gene expression in CHO cells was investigated (Figure 3C). Semi-quantitative RT–PCR showed that PEX1, PEX5, PEX13 and PEX14 are all induced to a comparable or greater extent than GST in cells treated with 1 mM H₂O₂ for 1 h.

Response of AtPEX1 to stress in whole plants

The PEX1 promoter was linked to the luciferase (LUC) gene to create transgenic plants (PEX1–LUC) that would report upon activation of PEX1 in response to stress. Figure 1 shows the fragment used. A translational fusion was made at the sequence DDWE, which is conserved in all known PEX1 genes. Depending on which start codon is used in vivo, a translational fusion with either 107 or 113 amino acids of the N-terminus of Pex1p fused to luciferase will be expressed.

All experiments were carried out on T2 progeny from two lines, 5ci and 1ci, which were recovered from independent transformations. Because the T2 plants of the two lines are segregating for the transgene, the absolute level of expression varies between plants and quantitative comparisons cannot be made. However, qualitative results are extremely consistent both within and between lines. No luciferase expression was seen in plants transformed with a promoter-less luciferase construct (data not shown). In untreated PEX1–LUC plants, PEX1–luciferase is mainly expressed in the meristems and young developing leaves, whereas there is little expression in mature leaves (controls in Figures 4A–C, 5B and 6B). These results suggest that PEX1 expression and peroxisome biogenesis are highest in actively dividing tissue and much lower in tissues where organelle maintenance (rather than biogenesis) is required. When PEX1–LUC plants were sprayed with 1 mM H₂O₂, PEX1–luciferase expression increased rapidly and dramatically throughout the plants (Figure 4A). This is consistent with the northern and RT–PCR results obtained with H₂O₂-treated wild-type plants (Figure 3A and B). To verify that increased PEX1 expression reflects increased peroxisome biogenesis, plants were subjected to two stress conditions in which peroxisomes have previously been described to proliferate, clofibrate treatment (Palma et al., 1991) and natural senescence (Pastori and del Rio, 1997). Incubation of PEX1–LUC plants with clofibrate resulted in a significant increase in the expression of PEX1–luciferase (Figure 4B). Compared with a young (20-day-old) plant, naturally senescent plants (60-day-old) also showed increased expression of PEX1–luciferase, which was no longer restricted to meristems and young tissue but was high in light yellow senescent leaves (Figure 4C).

Hydrogen peroxide induces PEX1 systemically

Hydrogen peroxide acts as a mobile signal (Alvarez et al., 1998; Karpinski et al., 1999). Opposite leaves of PEX1–LUC plants (Figure 5A, arrows) were incubated with a 1 mM H₂O₂ solution or a 1 mM H₂O₂ solution containing catalase (control) (Figure 5A). After 10 min incubation, a strong induction of PEX1–luciferase was detected throughout the treated leaf, but no photon emission was detected in the control leaf. By 30 min, induction of PEX1–luciferase in distant leaves, including the opposite (control) leaf, was observed.

Plant tissues rapidly break down exogenously applied H₂O₂ and it is difficult to estimate the resulting intracellular concentration. However, treatment of cell suspensions with the H₂O₂-regenerating system glucose (2.5 mM) and glucose oxidase (2.5 U/ml) (G/GO) initiates a steady and prolonged production of >5 µM H₂O₂. This mimics the level of H₂O₂ generated by the oxidative burst during pathogen attack (Delledonne et al., 1998). One leaf of a young PEX1–LUC plant was incubated with the G/GO H₂O₂-regenerating system (Figure 5B). PEX1–luciferase expression was induced locally and systemically within minutes of the start of the treatment, especially in vascular tissue. Note that photon emission from the treated leaf is much reduced, as it is submerged in G/GO solution.

The pathogen-induced hypersensitive response rapidly results in PEX1 induction

The ability to reproduce PEX1–luciferase induction with an authentic avirulent pathogen was tested. PEX1–LUC plants were inoculated in one leaf with Pseudomonas syringae pv. tomato, which induces an oxidative burst and hypersensitive response. An adjacent (control) leaf in the same plant was mock infiltrated. Photon emission in the inoculated and control leaves was monitored for 4 h immediately after inoculation (Figure 6A). Soon after infiltration of the pathogen, increased PEX1–luciferase expression was clearly seen in the treated leaf compared with expression in the control leaf and with the treated leaf before inoculation. Expression in the inoculated leaf was maximal at ∼2 h. Although the pathogen was inoculated at the tip of the leaf, gene expression was mainly localized around the vascular tissue, similar to when a leaf was incubated with 10 µM H₂O₂-regenerating solution (Figure 5B); however, the response to pathogen was localized rather than systemic.

PEX1 is induced by wounding

Plants respond to wounding and pathogen attack by distinct signalling mechanisms, although these pathways interact in a complex and incompletely understood manner (Malek and Dietrich, 1999). Photon emission was measured before wounding and then after excision of one or two leaves at the indicated times (Figure 6B–D). PEX1–luciferase expression was induced in the wounded leaf, but, as with pathogen treatment, the response remained localized. In the wounded leaves, PEX1– luciferase expression was observed mainly around the vascular tissues in the veins and the periphery of the leaf. In five independent experiments (two of which are shown) expression typically peaked between 45 and 120 min and then declined. Expression of PEX1 in the detached portion of the leaf continued to increase throughout the time course of the experiment, with expression increasing from the cut surface towards the tip of the leaf (data not shown). This may be the combined result of wounding and senescence responses.

Up-regulation of PEX genes by H₂O₂

In this paper we provide the first demonstration that peroxisome biogenesis is directly responsive to the major cellular stress signal H₂O₂. Peroxisomes are significant sources of active oxygen species as well as the site of important antioxidant enzymes and molecules. Consequently, peroxisomes are likely to play an important role in cellular redox balance. Steady-state mRNA levels of both plant and animal peroxisome biogenesis genes increase rapidly in response to 1 mM exogenous H₂O₂ treatment (Figure 3). PEX5 encodes the import receptor for PTS1-targeted proteins (Dodt et al., 1995), including catalase. PEX13 and PEX14 comprise the docking site for PTS1 and PTS2 import receptors and as such are central components of the peroxisomal import machinery (Gould et al., 1996; Albertini et al., 1997). PEX10 is a peroxisome membrane protein with a RING type zinc finger, which when mutated severely compromises the import of peroxisomal matrix proteins (Okumoto et al., 1998b).

AtPEX1 gene expression in whole plants in response to stress

Promoter–luciferase fusions are ideal for detecting rapid real time changes in gene expression with fidelity and high sensitivity (Millar et al., 1992). Using transgenic plants containing a PEX1–LUC transgene, exogenous application of 1 mM H₂O₂ resulted in the appearance of luciferase activity in the treated leaf within 10 min and the PEX1 promoter was activated systemically within 30 min (Figure 5A). This very rapid response argues strongly for activation of transcription of PEX1 being a direct consequence of the application of H₂O₂. The systemic induction is consistent with the known properties of H₂O₂ as a signal molecule (Levine et al., 1994). Sustained production of low micromolar concentrations of H₂O₂ by G/GO resulted in similar kinetics of induction of the PEX1 promoter (Figure 5B), demonstrating that this is a response to physiological levels of H₂O₂.

Infection with an avirulent pathogen results in the well-characterized production of H₂O₂ in vivo (Levine et al., 1994; Alvarez et al., 1998) and the induction of PEX1 by pathogen treatment is consistent with this (Figure 6A). This result demonstrates that peroxisome biogenesis is up-regulated when H₂O₂ is produced in a natural situation, thus it is a biologically meaningful response. However, the response is slower and not systemic. This may be because in plants time is required to generate the H₂O₂ and other mechanisms come into play to modulate the response.

Hydrogen peroxide is cytotoxic to the pathogen and to plant cells in the immediate vicinity, leading to a hypersensitive response. Protection of adjacent cells is achieved by induction of antioxidant defences, but as many of these are located in peroxisomes, elaboration of the peroxisome compartment might also be required early on to permit the uptake of newly synthesized antioxidant enzymes and molecules. The physiological importance of peroxisomal catalase as a sink for H₂O₂ has been demonstrated (Willekens et al., 1997). Of the three catalase genes in tobacco, Cat 2, which is preferentially expressed in vascular tissue (Willekens et al., 1994b), shows the highest induction by stress (Willekens et al., 1994a). Our results (Figures 4–6) show that PEX1– luciferase expression in response to various stresses is highest around vascular tissue. It is tempting to speculate that Cat 2 represents a stress-induced catalase that is imported into peroxisomes that are themselves elaborated in response to oxidative stress. In addition to protecting surrounding cells, peroxisomal antioxidants could also be responsible for turning off the H₂O₂ signal.

Wounding also produced a transient and localized response (Figure 6B). Wounding results in the induction of specific sets of genes, including those required for synthesis of jasmonates (Creelman and Mullett, 1997). Synthesis of jasmonic acid requires three cycles of (pre sumably) peroxisomal β-oxidation. Wounding also strongly induces one of the four acyl-CoA oxidases of Arabidopsis, AtACX1 (>10-fold induction within 1 h; Hooks et al., 1998). Either the fatty acids released from the phospholipids, jasmonic acid itself or the H₂O₂ produced by the wound-induced acyl-CoA oxidase could be responsible for induction of PEX1 upon wounding. The PEX1–LUC plants should be useful in testing these models.

A model for the control of peroxisome biogenesis by oxidative stress

Our results suggest a model whereby diverse stresses that generate H₂O₂ as a signalling molecule result in peroxisome proliferation via the up-regulation of components (PEX genes) required for biogenesis of the organelle and import of proteins. Previous work (Palma et al., 1991) has demonstrated that clofibrate treatment of pea plants results in elevated H₂O₂ and increased peroxisome number. The data presented here indicate that there is a causal relation ship between these two events and suggest a mechanism for it. Additionally, two other stresses, wounding and pathogen attack, which also generate H₂O₂ as a signalling molecule, were demonstrated to induce PEX1 expression. Our model would predict that other stress conditions which elevate H₂O₂, such as drought (Pei et al., 2000), osmotic stress (Guan et al., 2000) and excess light (Karpinski et al., 1999), would also result in PEX gene induction and peroxisome proliferation. Peroxisome number and morphology has been reported to increase as a function of light intensity (Ferreira et al., 1989).

Induction of PEX genes by H₂O₂ is not confined to plants. Expression of all mammalian PEX genes tested was significantly increased by exposure to H₂O₂, suggesting that this may be a common mechanism for dealing with oxidative stress. Induction of peroxisome biogenesis by this general oxidative stress signal fits nicely with the presence of so many antioxidant enzymes within peroxisomes and appears to be an evolutionarily ancient response of cells to stress.

Biological material

Arabidopsis thaliana ecotype Colombia was used for all experiments. Wild-type seeds were sown on 1/2 MS, 0.7% agar plates containing 500 µg/ml carbenicillin (Melford Laboratories), and 2.5 µg/ml amphotericin B (Sigma) and incubated in 16 h light, 8 h dark at 18°C. Transgenic seeds were sown in pots containing a 9:1 compost:sand mixture and grown in a Sanyo growth chamber at 18°C with 9 h light (240 µmol/m²/s), 15 h dark. Experiments with transgenic plants were carried out with T2 seeds from two independent transgenic lines (5ci and 1ci) and were repeated between two and six times. CHO cells were obtained from the European Collection of Cell Cultures (Salisbury, Wiltshire) and were cultured in Ham’s F12 nutrient mix (Gibco BRL, Paisley, UK) supplemented with 2 mM l-glutamine, 10% (v/v) fetal bovine serum, 100 U/ml penicillin and 100 µg/ml streptomycin at 37°C with 5% CO₂.

Isolation of a PEX1 cDNA clone

A tBLASTn search with S.cerevisiae Pex1p identified EST T43478 as a possible Arabidopsis orthologue (Kaplan, 1996) and a BAC clone MAH20 (DDBJ/EMBL/GenBank accession No. AB006697) from chromosome 5 was found that showed a highly significant identity (P = 5.6 × 10^–160) to the EST. An 899 bp genomic DNA probe was amplified by PCR and used to screen a fractionated Arabidopsis 3 day seedling hypocotyl cDNA library (CD4-16, size range 3–6 kb; ABRC DNA Stock Centre, Ohio State University). The longest clone (3.5 kb) was sequenced commercially (MWG Biotech). A partial cDNA clone of 215 bp, which extends the sequence 5′, was isolated from a 2 day seedling library.

Isolation of the PEX1 promoter

A fragment of 1436 bp comprising 441 bp of coding sequence and 995 bp of 5′ sequence was amplified by PCR (Figure 1). Clones from three independent PCRs were sequenced and verified against the genomic sequence. The fragment was cloned as a PstI–NcoI fragment into pGRT10 (pUC18 into which the luciferase gene had previously been inserted) to produce pPEX1-GRT. The PEX1–luciferase cassette was excised with NotI (blunted) and SacI and inserted into SmaI–SacI-digested pBI101. To produce a promoterless control, luciferase was excised from pGRT10 with HindIII and SacI and ligated with HindIII–SacI-digested pBI101. Transformation of Arabidopsis was according to the protocol of A.Bent (University of Illinois at Urbana, http://www.cropsci.uiuc.edu/~a-bent/protocol.htm).

RNA analysis

For H₂O₂ treatments, leaves from young plants were detached and immediately submerged in 1 mM H₂O₂ ± excess catalase (Sigma) for 1 h at room temperature, then frozen in liquid nitrogen and stored at –80°C. CHO cells were treated with 1 mM H₂O₂ solution in situ for 1 h at room temperature with gentle shaking. RNA was extracted using the Qiagen RNeasy Midi kit (catalogue no. 75142). Total RNA (10 µg) was separated in a formaldehyde gel and transferred to Hybond NX membrane (Amersham Pharmacia Biotech UK). Probes were amplified by PCR from the corresponding cDNA clone (PEX1); from a 2 day seedling cDNA library and sequenced (PEX5); and from an EST (GST, clone ATTS1553; ABRC, Ohio State University). Probes were labelled using a Rediprime kit and unincorporated label removed using Nick columns (both from Amersham Pharmacia Biotech UK). Hybridization was as described by Hooks et al. (1999). RT–PCRs were carried out using the Reverse-iT One Step kit from the AB gene following the manufacturer’s instructions with equal amounts of RNA from control and H₂O₂-treated samples.

Plant treatments

For H₂O₂ treatments of whole plants a solution of H₂O₂ (1 mM in 0.9% NaCl, 10 mM Tris–HCl pH 7.5) was sprayed from above. For leaf incubations, one leaf was submerged either in 50 ml of 1 mM H₂O₂ in 0.9% NaCl, 10 mM Tris–HCl pH 7.5 for 15 min or in a 5–10 µM H₂O₂ solution and an H₂O₂ regenerating system consisting of glucose (2.5 mM) and glucose oxidase (2.5 U/ml). For clofibrate treatment, 25 µl of a 100 mM solution were placed on opposite leaves of the plant. For pathogen inoculation and wounding experiments, 3- to 5-week-old plants that had not initiated flowering were used. Pseudomonas syringae pv. tomato (Pst) strain DC3000 carrying the avrRpm1 gene was grown overnight from a plate in King’s B Medium containing rifampicin (50 µg/ml) and kanamycin (25 µg/ml), washed twice with 10 mM MgCl₂, and ∼5–10 µl of a 10⁶–10⁷ c.f.u./ml suspension in 10 mM MgCl₂ were gently inoculated into the underside of an intact leaf using a 1 ml sterile plastic syringe without a needle. Control inoculations were carried out with 10 mM MgCl₂. For wounding experiments, one or two leaves were cut with scissors. The wounding and pathogen experiments were conducted with plants that had been kept in the dark overnight. Wounding and G/GO treatments were carried out in the camera to avoid exposing the plants to light between recording of the background image and data capture.

In vivo imaging of luciferase bioluminescence

An intensifier camera (VIM) and photon counting image processor (ARGUS-50) (Hamamatsu Photonics UK) were used for all experiments. Arabidopsis plants were sprayed with 5 mM d-luciferin solution (Analytical Bioluminescence Laboratories, San Diego, CA) for 12 h (Millar et al., 1992) and then again 20 min before photon emission measurements were taken. Plants were left within the camera case for at least 5 min in complete darkness to allow auto-luminescence to decay before measurements were taken. A reference measurement (control) was always taken prior to treatments.