Autotoxicity mechanism of Oryza sativa: transcriptome response in rice roots exposed to ferulic acid

Background Autotoxicity plays an important role in regulating crop yield and quality. To help characterize the autotoxicity mechanism of rice, we performed a large-scale, transcriptomic analysis of the rice root response to ferulic acid, an autotoxin from rice straw. Results Root growth rate was decreased and reactive oxygen species, calcium content and lipoxygenase activity were increased with increasing ferulic acid concentration in roots. Transcriptome analysis revealed more transcripts responsive to short ferulic-acid exposure (1- and 3-h treatments, 1,204 genes) than long exposure (24 h, 176 genes). Induced genes were involved in cell wall formation, chemical detoxification, secondary metabolism, signal transduction, and abiotic stress response. Genes associated with signaling and biosynthesis for ethylene and jasmonic acid were upregulated with ferulic acid. Ferulic acid upregulated ATP-binding cassette and amino acid/auxin permease transporters as well as genes encoding signaling components such as leucine-rich repeat VIII and receptor-like cytoplasmic kinases VII protein kinases, APETALA2/ethylene response factor, WRKY, MYB and Zinc-finger protein expressed in inflorescence meristem transcription factors. Conclusions The results of a transcriptome analysis suggest the molecular mechanisms of plants in response to FA, including toxicity, detoxicification and signaling machinery. FA may have a significant effect on inhibiting rice root elongation through modulating ET and JA hormone homeostasis. FA-induced gene expression of AAAP transporters may contribute to detoxicification of the autotoxin. Moreover, the WRKY and Myb TFs and LRR-VIII and SD-2b kinases might regulate downstream genes under FA stress but not general allelochemical stress. This comprehensive description of gene expression information could greatly facilitate our understanding of the mechanisms of autotoxicity in plants.


Background
Monoculture of crops leads to decreased growth and yield in the next season, with autotoxicity the major culprit [1][2][3]. Autotoxicity occurs when a plant releases toxic chemical substances into the environment that inhibit germination and growth of conspecific plants [4]. Recently, an increasing number of reports have provided evidence for the role of autotoxicity in replant failure and soil sickness [1]. Autotoxicity is a common problem in continuous monocropping of rice [2] because decomposing rice straw is left in fallow fields [5]. A range of secondary metabolites in rice straws, such as phenolic acids [6] and a few flavones and terpenoids [7], are potent autotoxins.
Phenolic compounds are common in soils. Whitehead [8] reported that the concentration of phenolic compounds in rhizosphere soil solution may reach 90 ppm. Various phenolic compounds such as ferulic acid (FA), o-hydroxy phenyl acetic acid, and p-coumaric acid have been isolated from decomposing rice residues in soil [5].
These compounds inhibit the growth of rice seedlings in the order of FA > p-coumaric acid > o-hydroxy phenyl acetic acid [9]. Exposure of plant roots to FA reduces water use [10], inhibits foliar expansion [11] and root elongation [12], and decreases nutrient uptake [13][14][15]. Further, FA exposure rapidly depolarizes root cell membranes, causing a generalized increase in membrane permeability, inducing lipid peroxidation and affecting certain enzymatic activities [16][17][18]. Ferulic acid may be esterified with cell wall polysaccharides, be incorporated into lignin structures, or form bridges that connect lignin with wall polysaccharides, thus resulting in cell wall rigidity and restriction of cell growth [19,20]. Ferulic acid affects cell wall-bound peroxidase (POD) and phenylalanine ammonia-lyase (PAL) activities, lignin content, and root growth in seedlings [21].
Several reports demonstrated that autotoxins induce oxidative stress in plants [22,23]. Reactive oxygen species (ROS) play a vital role in the plant defense against stresses and in cell growth and development [24,25]. Low concentrations of ROS, as a signal, can lead to repair of cellular damage, but high levels can lead to programmed cell death [26]. Calcium is a crucial regulator of growth and development in plants [27]. ROS-activated calcium channel activity is required during the growth of cells in the elongation zone of the root [28].
Both allelopathy and autotoxicity play important roles in regulating plant biodiversity and productivity [3]. Autotoxins can impact many physiological and biochemical reactions in plants such as rice, alfalfa, cucumber, tomato, corn, wheat, sugarcane [1,23]. The potential mechanisms underlying autotoxicity have been explored in alfalfa and cucumber [22,29]. In alfafa, cinnamic acid is a phenolic acid and the major autotoxin in leaves and root exudates [30]. In cucumber, autotoxins can inhibit the membrane H + -ATPase activity that drives the uptake of essential ions, other solutes and water [22]. However, our knowledge of an autotoxicity mechanism is poorly understood. Transcriptional profiling experiments using microarrays are being conducted to examine the effects of natural phytotoxins on the plant transcriptome [31]. Microarray analyses were used to analyze gene expression profiles of plants exposed to the allelochemicals 2(3H)benzoxazolinone [32], fagomine, gallic acid, rutin [33], 3-(3',4'-dihydroxyphenyl)-L-alanine [34], and juglone [35].
Rice (Oryza sativa L.) is a model for genomic research into the responses of monocot species to environmental stresses. In this study, we used FA as a rice-model autotoxin and used microarray assay to assess alterations in rice root gene expression induced by the autotoxin. We discuss the possible involvement of reactive oxygen species (ROS) and calcium in allelochemical signal transduction pathways. These data significantly expand on previous studies examining plant transcriptional responses to allelochemicals and provides a foundation for elucidating the autotoxicity mechanism of O. sativa, particularly the phytotoxic effect of decomposing rice residues in soil.

Results
Effect of FA on growth and root architecture of rice To select an appropriate concentration of FA for stress treatments, we conducted a dose-response analysis of rice root growth 3 days after FA treatment ( Figure 1A). Compared with the control, 25 ppm FA significantly reduced root growth. With 50 ppm FA, root growth was about half of the control growth, and with 200 ppm, growth was almost completely inhibited.
Compared with the control, 50 ppm FA inhibited crown root, lateral root and root hair formation. Both the number and length of lateral roots and root hairs were significantly reduced with 50 ppm FA as compared with the control (Table 1).

Ferulic acid rapidly induced ROS and calcium accumulation in rice roots
To determine whether FA treatment induced ROS production, we labeled roots with the ROS-sensitive dye CM-H 2 DCFDA ( Figure 1B) or nitroblue tetrazolium (Additional file 1: Figure S1), then treated them with 50 ppm FA for 1 or 3 h. Ferulic-acid stress significantly increased the levels of dihydrodichlorofluorescein (DCF), and thus ROS, in roots ( Figure 1B, Additional file 2: Figure S2). To determine whether FA treatment induced calcium accumulation, we used a calcium indicator, Oregon green 488 BAPTA-1, before FA treatment. Calcium level was significantly increased in root tip regions with 50 ppm FA treatment for 1 or 3 h ( Figure 1C and Additional file 2: Figure S2).

Effect of FA on lipid peroxidation
Ferulic-acid-induced oxidative damage of roots was positively confirmed by Schiff 's staining in the meristem and elongation zone of roots (Additional file 3: Figure  S3). Ferulic-acid-induced root oxidative damage was measured by LOX activity with non-denaturing PAGE. We detected 3 LOX isozymes in rice roots treated with 50 ppm FA for 3, 6, 12, and 24 h (Additional file 3: Figure S3).

Expression profiling by microarray assay
To identify genes and biological pathways associated with FA toxicity and tolerance in rice roots, we used large-scale expression profiling. RNA samples were collected from root tips early (1 and 3 h) after FA treatment to examine rapid changes in global patterns of gene expression. We pooled RNA isolated from the two short (1 and 3 h) FA exposures to maximize gene discovery. Mechanisms of adaptation after long-term (24 h) FA exposure are important, but the physiological and metabolic parameters measured after long treatment periods might be distorted by the severe toxic effects of FA. We aimed to understand the primary response to FA exposure as opposed to responses to nonspecific cellular damage.
We performed microarray assays with RNA extracted from roots treated with 50 ppm FA after short (pooled from 1-and 3-h treatments) and long (24 h) exposure. This FA level is comparable to that found in rice-field soils [8,36]. In all, 1,204 genes were responsive to short FA exposure and 176 to long exposure. After short FA treatment, 972 genes were upregulated (FDR < 0.1, fold change ≥ 2) and 232 were downregulated (FDR < 0.1, fold change ≤ 0.5) (Additional file 4: Table S1).
These observations were further supported by comparison of metabolism genes with use of MapMan. The genes encoding enzymes related to detoxification were cytochrome P450, UDP glycosyltransferases, and glutathione-S -transferases ( Figure 2A). RT-PCR validated the microarray findings (Additional file 6: Figure S4).

Expression profiles of root architecture-related genes
To investigate the involvement of root architecture related genes in FA-induced stress, we analyzed the global expression profiles of genes related to 3 such gene families (Table 3). In total, 3 of the 18 root architecture related genes were slightly downregulated by FA (FDR < 0.1). FA repressed the expression of two lateral-root genes (ARF-16, Os06g0196700, downregulated 1.7-fold; OsCel9C, Os05g0212300, downregulated 1.6-fold) and one root-hair-related gene (OsCSLD1, Os10g0578200, downregulated 1.5-fold) in rice roots (Table 3).

Expression profiles of transporter genes
In the rice genome, transporter families are grouped by mode of transport and energy-coupling mechanism into four types: ATP-dependent transporters, secondary transporters, ion channels, and unclassified transporters with unknown mechanisms of action. Among 1,286 transporter-related genes, 1,113 were present on our arrays, and 64 were significantly upregulated with FA treatment ( Table 5, Additional file 9: Table S5). Nearly all of the transporters responding to FA were ATPdependent and secondary transporters. Transporters with changed expression were 17 of the 130 ATP-binding cassette (ABC) transporters and three of the P-type ATPase (P-ATPase) transporters. The major facilitator superfamily (MFS) is the largest family of secondary transporters in the rice genome. Ferulic-acid treatment upregulated nine MFS genes and downregulated two. Transcripts for five proton-dependent oligopeptide transporter (POT) genes and five amino acid/auxin permease (AAAP) genes were upregulated. In addition, four of 123 drug/metabolite transporter (DMT) genes belonging to secondary transporters were upregulated by FA treatment.
These observations were further supported by comparison of metabolism genes by use of MapMan. Genes encoding ATP-binding cassette-type and AAAP transporters were differentially regulated in the early (1 and 3 h) response to FA ( Figure 2B). MapMan analysis revealed that AAAP transporters were significantly upregulated by FA treatment.

Expression profiles of phytohormone-related genes
Among 324 phytohormone-related genes, 297 were present on our arrays, and 25 were significantly upregulated with FA treatment ( Table 6, Additional file 10: Table S6). One jasmonic acid (JA) biosynthesis gene, OsAOS2 (Os03g0767000) and six JA signaling genes (Os03g0180900, Os10g0392400, Os03g0402800, Os03g01   81100, Os03g0180800, and Os09g0439200) were upregulated by FA exposure; none were downregulated during the same time of exposure. MapMan analysis revealed that ethylene (ET) synthesis and signaling genes were significantly upregulated by FA treatment ( Figure 2C).

Expression profiles of signaling genes and TFs
Perception and transmission of stress signals are important aspects of the plant response to environment stress. Protein kinases are crucial in these signaling pathways. The activation of signal transduction pathways connects the actions of protein kinases, TFs and the downstream stress-responsive genes. In total, 51 protein kinase genes were upregulated by FA, and 16 were downregulated ( Figure 3A, Additional file 11: Table S7). Nearly all of the FA-responsive kinases were associated with the receptor-like kinase (RLK) family. In total, 40 RLK family genes were significantly upregulated and 15 were downregulated after short and long FA exposure. The leucinerich repeat VIII (LRR-VIII) and receptor-like cytoplasmic kinases VII (RLCK-VII) subfamilies of the RLK family were significantly upregulated with FA treatment. We found 107 TFs significantly regulated by FA: 85 were significantly upregulated and 22 downregulated after short and long exposure. Transcription factors regulated by FA stress predominantly belong to the APETALA2/ET response factor (AP2/ERF), MYB, WRKY and Zinc-finger protein expressed in inflorescence meristem (ZIM) families ( Figure 3B, Additional file 12: Table  S8). From rice genome sequence data, 164, 129, 100 and 18 genes have been identified for the AP2/ERF, MYB, WRKY and ZIM families, respectively. In our rice roots, FA induced 14 AP2/ERF, 11 MYB, 17 WRKY and 6 ZIM families.
Transporters, TFs, and protein kinases specifically altered by FA and juglone We compared transporters, TFs, and protein kinases regulated by exposure to FA and to the ROS-generating allelochemical juglone (Figure 4). Genes encoding AAAP transporters responded relatively specifically to FA (Additional file 13: Table S9). Comparison of the TFs induced by juglone after short FA exposure revealed that only half of the genes (48 of 84) reported in our previous study [35] showed changed expression in this study (Additional file 14: Table S10). The WRKY and Myb TFs responded significantly to FA stress. Comparison of the protein kinase genes induced by juglone revealed that the LRR-VIII and SD-2b families responded significantly to FA stress (Additional file 15: Table S11).

Discussion
Autotoxicity is intraspecific allelopathy and plays an important role in plant growth inhibition and structuring plant communities [1]. Here, we observed that treatment with 50 ppm of the autotoxic chemical FA inhibited rice root length by 50% (Figure 1). Plant growth as well as response to stress is controlled by phytohormones [38]. Our microarray assay revealed changed expression of ET and JA biosynthesis and signaling genes in rice roots with FA treatment (Table 6). Xu et al. [39] reported that the combination of ET and JA synergistically induced the expression of defense genes in plants. Hua and Meyerowitz [40] and Staswick et al. [41] reported that JA and ET treatment inhibited plant root elongation. Our results suggest that these two hormones may be involved in FA-induced inhibition of root growth in rice. In addition, we found that FA inhibited crown root, lateral root and root hair formation. FA repressed the expression of two lateral-root genes (ARF-16, Os06g0196700; OsCel9C, Os05g0212300) and one root-hair-related gene in rice roots (OsCSLD1, Os10g0578200) ( Table 3). Examination of the RiceXPro database revealed that JA repressed the expression of these genes (ARF-16, OsCel9C and OsCSLD1) [42]. Therefore, our results suggest that JA may involve in FA-induced morphogenic response in roots by regulating these root architecture-related genes. In general, cell walls become lignified when cell expansion decreases or when the cell is under stress [43]. Previous study indicated that lignification may be an important step in root growth reduction in FA-stressed soybean [21]. Our FA treatment upregulated genes involved in the cell-wall macromolecule metabolic process (GO:0044036, FDR 6.10E-13), particularly cell-wall reassembly. The expression of cell-wall-related genes, most notably the expansins, was enriched by FA treatment. Cosgrove found that expansins are a group of wall proteins that induce wall stress relaxation and extension [44]. Increased expansin gene activity may be involved in stress relaxation in FA-treated roots.
Increased ROS levels are an important component of the stress induced by allelochemicals [45]. Ferulic acid modifies various oxidative reactions in vitro by acting as a substrate, activator or inhibitor depending on the concentration [46,47]. Reactive oxygen species are toxic to plant tissues and can trigger cell growth inhibition and cell death. In addition, they may act as signal molecules involved in triggering tolerance against various environmental stresses. In this study, FA rapidly induced ROS production in rice roots. Ferulic acid-induced lipid peroxidation of roots was positively confirmed by Schiff 's reagent staining (Additional file 3: Figure S3). We found a steady increase in LOX activity in response to FA (Additional file 3: Figure S3). Thus, FA may increase ROS accumulation, lipid peroxidation, and LOX activity to affect cell integrity in rice roots and contribute to FAinduced root growth inhibition.
Many studies have provided evidence that ROS signaling is integrated with calcium signaling networks in plants. Saijo et al. [48] and Martín and Busconi [49] demonstrated rapid increase in cytoplasmic calcium concentrations in plant cells in response to multiple stress stimuli. The change in cytoplasmic calcium concentrations is critical for activating various defense responses [50]. We found that FA increased calcium levels in rice roots. Thus, ROS and calcium may act as early second messengers in the transcriptional activation of an
a The fold change in expression of each gene after FA treatment was calculated by the mean from 3 biological replicates and false discovery rate <0.1 is shown with a asterisk.
array of defense-related genes in rice roots under FA stress. Conjugated forms of xenobiotics can be recognized by specific membrane-associated transporters in the final detoxification phase [32]. Our GO analysis notably revealed the term "primary active transmembrane transporter activity". We found 64 membrane-transporter-like sequences induced by FA, including 17 putative ABC, nine MFS, and five AAAP transporters. In plants, ABC and MFS transporters represent different multidrug efflux protein superfamilies associated with resistance to xenobiotics [32]. The ABC transporters facilitate the movement of glutathionylated toxins and other substrates across biological membranes [51]. We found 17 and three ABC transporter genes upregulated by short and long FA exposure, respectively. Thus, expression of ABC transporters, which work in conjunction with other detoxifying systems, was found primarily with early stages of FA stress. The AAAPs are efficient transporters of proline and betaine [52] that accumulate in higher plants under stress conditions such as drought, salinity, extreme temperatures, UV radiation, and heavy metals [53,54]. Previous reports have demonstrated a positive relationship between proline and betaine accumulation and plant stress tolerance [55,56]. Our observed induction of AAAPs by FA indicates their possible involvement in plant tolerance to autotoxin stress.
Protein kinases are important signaling molecules in the plant response to environment stress. Multiple plant RLK members are involved in the stress response [57][58][59]. Among 40 RLK genes we found upregulated with FA treatment, LRR-VIII and RLCK-VII subfamilies were identified as significantly participating in transcriptional regulation (Additional file 11: Table S7). The involvement of LRR-VIII and RLCK-VII in stress responses was previously reported [58]. Thus, differential expression of a number of transmembrane receptor kinases with FA exposure suggests that multiple receptors belonging to different families may have unique regulatory mechanisms.
Responses to abiotic stresses require the production of important regulatory proteins such as TFs to mediate the expression of downstream stress-responsive genes. We found that the major TFs, AP2/ERF, MYB, WRKY, and ZIM, were overrepresented in the response to FA. The AP2/ERF, MYB, and WRKY TFs have been isolated from different plants and are important candidates for the stress tolerance response; in rice, the overexpression of AP2/ERF, MYB, and WRKY conferred significant tolerance to abiotic stresses [60][61][62][63]. Transcription factors of ZIM have been intensively investigated because of the role of these proteins as key regulators of the jasmonate hormonal response in Arabidopsis and rice [64]. Here, we found that FA upregulated six ZIM genes. Overexpression of ZIM-3 (Os03g0180800), a stress-inducible gene, was      found to significantly increase tolerance to salt and dehydration stresses [64]. The observed induction of AP2/ ERF, MYB, WRKY, and ZIM TFs during FA treatment indicates their possible involvement in plant resistance to autotoxin stress. Reactive oxygen species are secondary messengers for the activation of specific TFs. We found that FA induced ROS production. Therefore, we compared the set of our FA-regulated TFs to those regulated by exposure to juglone, an ROS-generating allelochemical [35]. Our results suggest that WRKY and Myb TFs and LRR-VIII and SD-2b kinases might regulate downstream genes under FA stress but not general allelochemical stress ( Figure 4). Moreover, 64 transporters were upregulated by FA, but only 31 transporters were upregulated by juglone. The number of upregulated genes encoding transporters was more under FA than juglone stress. Especially, the AAAP transporter family was regulated significantly by FA stress but not by juglone (Figure 4). The AAAPs are efficient transporters of osmoprotectants such as proline, glycinebetaine and gamma-aminobutyric acid [52] that accumulate in higher plants under stress conditions. This observation could be related to detoxification of the autotoxin in rice roots. The AAAP transporters may play an important role in the FA-triggered autotoxicity mechanism.

Conclusions
FA may have a significant effect on inhibiting rice root elongation through ET and JA gene regulation. Detoxification enzymes such as cytochrome, GST, and ROS scavengers are involved in protecting against FA toxicity. Moreover, proteins involved in regulatory functions and signal transduction, including TFs, calcium-regulated proteins, and various protein kinases, play important roles in the response to FA stress ( Figure 5). Future studies with rice mutants or overexpressors with altered expression of the genes identified in this work will be helpful to elucidate their biological significance and clarify new pathways involved in toxicity and tolerance to FA.

Plant materials
Rice plants (O. sativa L. cv. TN-67) were grown as previously described [65]. Rice seedlings were exposed to FA (25 to 200 ppm) for 1 to 24 h. Control plants were treated with water in parallel for the indicated times.

Analysis of growth
Rice seeds were surface-disinfected with 2.5% (v/v) sodium hypochlorite (Katayama, Osaka, Japan) for 15 min, then thoroughly washed in distilled water. Seeds were placed in 9-cm Petri dishes containing 20 ml distilled water and left at 37°C in the dark. After 2 d of incubation, uniformly germinated seeds were transferred to Petri dishes with filter paper discs (Advantec, Tokyo) moistened with 10 ml distilled water. Each Petri dish contained 15 germinated seeds grown at 27°C in the dark for 3 d. To determine the length of root hairs, the length of the 20 longest root hairs from the root hair zone of seminal roots was measured.   were used for microarray assay and RT-PCR.

Microarray preparation and analysis
Six-day-old rice seedlings were exposed to 50 ppm FA for short (1 and 3 h) or long (24 h) durations, then RNA was isolated from root tips to examine rapid changes in global patterns of gene expression. We pooled RNA from the two short exposures to maximize gene discovery. RNA from water-treated (control) and FA-treated roots was used with the Agilent Rice Oligo microarray (4 × 44 K, custom-made; Agilent Technologies, Palo Alto, CA, USA) for RNA labeling and microarray hybridizations involved 3 biological replicate samples.
For the microarray assay, 0.5 μg total RNA was amplified by use of the Fluorescent Linear Amplification Kit (Agilent) and labeled with Cy3-CTP (control samples) or Cy5-CTP (FA-treated) (CyDye, PerkinElmer, Norwalk, CT, USA) during the in vitro transcription process. In total, 0.825 μg Cy-labeled cRNA was fragmented to a mean size of about 50-100 nt by incubation with fragmentation buffer (Agilent) at 60°C for 30 min. The fragmented-labeled cRNA was then pooled and hybridized to the Rice Oligo DNA Microarray 44 K RAP-DB (G2519F#15241; Agilent) at 60°C for 17 h. After a washing and blow-drying with a nitrogen gun, microarrays were scanned with use of an Agilent microarray scanner at 535 nm for Cy3 and 625 nm for Cy5. Scanned images were analyzed with use of Feature Extraction v9.5.3 (Agilent), with LOWESS normalization.
Signal intensities were extracted with use of Feature Extraction v9.5.3. For statistical analysis, we excluded genes with signal intensities < 100 in all experiments. Significant differences from 0 were identified by use of t test with GeneSpringGX11 (Agilent). The Benjamini-Hochberg false discovery rate (FDR) method was used to  obtain P-values that were corrected for multiple testing. The fold change in expression of each gene after FA treatment was calculated by the mean from 3 biological replicates. Genes upregulated by FA treatment by more than two-fold (cutoff by FDR < 0.1) were extracted. Each probe was considered an individual gene and annotated according to the Rice Annotation Project Data Base (RAP-DB; http://rapdb.dna.affrc.go. jp/; Rice Annotation Project 2007, 2008). The dye swap was not included. Three biological replicates were performed with 3 independent microarray slides for both short-and long-term FA treatments. Total RNA control samples were labeled with Cy3, and total RNA experimental samples (FA treatment) were labeled with Cy5.

MapMan display
The averaged signals for a given treatment were expressed relative to those for control samples, converted to a log2 scale and displayed by use of MapMan v3.5.1 [75]. O. sativa mapping files were imported into MapMan. Rice genes represented on the Rice Oligo DNA Microarray were organized by BINS and sub-BINS for display on the schematic map of the transport overview. Gene expression was analysed by the Wilcoxon Rank Sum test with uncorrected p value.

Semi-quantitative RT-PCR
Total RNA was isolated from root tissues treated with 50 ppm FA for 3, 12, or 24 h by use of the RNeasy Plant Mini kit (QIAGEN) and purified with the RNase-Free DNase Set (QIAGEN). Primer sequences are in Supporting Information (Additional file 16: Table S12). The number of PCR cycles in the experiments was adjusted to the optimal conditions. The data was shown on the basis of at least three biological replicates. Amplicons were analyzed by 1% agarose gel electrophoresis, and PCR products were sequenced.

Histochemical analyses and in-gel enzyme analyses
Histochemical detection of lipid peroxidation involved use of Schiff's reagent [76]. In brief, freshly harvested rice roots were stained with Schiff's reagent for 60 min, which detects aldehydes originating from lipid peroxides. Then roots were rinsed with potassium sulphite solution (0.5% [w/v] K 2 S 2 O 5 prepared in 0.05 M HCl) and maintained in the solution. The isozymes of lipoxygenase (LOX) were separated on discontinuous polyacrylamide gels (stacking gel 4.5%, separating gel 10%) under nondenaturing and non-reducing conditions. Proteins were electrophoretically separated at 4°C and 80 V in the stacking gel, then 120 V in the separating gel. Isozymes of LOX were visualized as described [77].