The Zinc Finger Protein Zat 12 Is Required for Cytosolic Ascorbate Peroxidase 1 Expression during Oxidative Stress in Arabidopsis *

Cytosolic ascorbate peroxidase 1 (Apx1) is a key H2O2 removal enzyme in plants. Microarray analysis of Apx1deficient Arabidopsis plants revealed that the expression of two zinc finger proteins (Zat12 and Zat7) and a WRKY transcription factor (WRKY25) is elevated in knock-out Apx1 plants grown under controlled conditions. Because mutants lacking Apx1 accumulate H2O2, we examined the correlation between H2O2 and the expression of Zat12, Zat7, WRKY25, and Apx1. The expression of Zat12, Zat7, and WRKY25 was simultaneously elevated in cells in response to oxidative stress (i.e. H2O2 or paraquat application), heat shock, or wounding. In contrast, light or osmotic stress did not enhance the expression of these putative transcription factors. All stresses tested enhanced the expression of Apx1. Transgenic plants expressing Zat12 or Zat7 could tolerate oxidative stress. In contrast, transgenic plants expressing WRKY25 could not. Although the expression of Zat12, Zat7, or WRKY25 in transgenic plants did not enhance the expression of Apx1 under controlled conditions, Zat12-deficient plants were unable to enhance the expression of Apx1, Zat7, or WRKY25 in response to H2O2 or paraquat application. Zat12-deficient plants were also more sensitive than wild type plants to H2O2 application as revealed by a higher level of H2O2-induced protein oxidation detected in these plants by protein blots. Our results suggest that Zat12 is an important component of the oxidative stress response signal transduction network of Arabidopsis required for Zat7, WRKY25, and Apx1 expression during oxidative stress.

Plants are sessile organisms that evolved a complex and specialized network of regulatory genes to control their response to changes in environmental conditions. It is likely that many of these regulatory genes were initially created by gene duplication and that they later acquired roles specifically related to individual pathways or stresses as well as their combination (1,2). Different members of gene families, such as WRKY and other zinc finger proteins (72 WRKY genes and over 600 zinc finger proteins in Arabidopsis; Ref. 3), MYB transcription factors (133 genes in Arabidopsis; Ref. 4), and heat shock transcription factors (21 genes in Arabidopsis; Ref. 5), were found to control and regulate diverse processes in plants ranging from development to response to biotic or abiotic stresses (1)(2)(3)(4)(5).
The different regulatory networks of plants are also involved in modulating the production and scavenging of reactive oxygen species (ROS) 1 in cells. These toxic intermediates of oxygen reduction not only control different plant responses to environmental and developmental cues but also potently inhibit essential metabolic pathways and may lead to cell death (6 -9). Although a number of different enzymes and proteins produce or scavenge ROS in cells, little is known about how the different regulatory networks of plants control these enzymes and proteins and modulate the steady-state level of ROS (8 -10). The steady-state level of a number of different transcripts encoding transcription factors such as MYB, WRKY, heat shock transcription factors, and different zinc finger proteins is elevated in plants in response to different forms of ROS-induced stress (11)(12)(13)(14). However, genetic evidence supporting a direct regulatory role for these transcripts was only presented for two zinc finger proteins, Lsd1 and Lol1, which were recently found to mediate ROS signals and control programmed cell death in Arabidopsis (6), and for heat shock transcription factor 3, which was shown to enhance cytosolic ascorbate peroxidase (Apx) expression in the absence of stress (15) We are studying the response of plants to overaccumulation of ROS in cells (i.e. oxidative stress; Ref. 9). Our goal is to identify and characterize the transcription factor network that controls the response of plants to oxidative stress. To dissect and study the ROS signal transduction network of plants, we are using knock-out plants deficient in key ROS-scavenging enzymes (13,14). These plants provide an ideal experimental system to study plant responses to ROS accumulation, because they accumulate ROS and activate multiple defense mechanisms in the absence of externally applied stimuli such as stress, ROS, or ROS generators. Moreover, the ROS that accumulate in these mutants are ROS naturally produced in cells at the different cellular ROS-producing sites and not externally applied ROS that may activate additional signaling pathways, including pathogen or abiotic stress response pathways (13,14). Knock-out plants deficient in cytosolic Apx1 are of particular interest. They maintain a high steady-state level of H 2 O 2 in cells and activate ROS defense mechanisms when grown under controlled conditions (13). These plants are also altered in their growth, flowering time, and stomatal responses and display an augmented induction of heat shock proteins and catalase in response to light stress.
Microarray analysis of knock-out Apx1 plants grown under controlled conditions revealed that the expression of at least two different zinc finger proteins (Zat7 and Zat12), a putative WRKY transcription factor (WRKY25), and a number of heat shock transcription factors is elevated in these plants (13). The expression of Zat12 is also elevated in cultured Arabidopsis cells in response to H 2 O 2 application (12) and in mature Arabidopsis plants in response to cold stress, wounding, or high light stress (Refs. 16 -18; stresses that result in ROS accumulation in cells). No direct link was reported between Zat12 expression and the expression of different ROS-scavenging transcripts such as those encoding Apx1. This finding stands in contrast to the established relationship between Apx1 expression and heat shock transcription factors (13,15). The elevated expression of Zat12, Zat7, and WRKY25 in Apx1-deficient plants suggests a link between H 2 O 2 accumulation, the expression of these putative regulatory genes, and Apx1 expression. In this study we examined the relationship between Zat12, Zat7, and WRKY25 expression, oxidative stress, and Apx1 expression. Our results suggest that Zat12 is essential for Zat7, WRKY25, and Apx1 expression during oxidative stress and that Zat12, Zat7, and WRKY25 are linked to H 2 O 2 stress in plants.

EXPERIMENTAL PROCEDURES
Plant Material and Growth Conditions-Arabidopsis thaliana (cv. Columbia and WS) plants were grown in growth chambers (Percival E-30) under controlled conditions (21-22°C, 18 h or constant light cycle, 100 mol m Ϫ2 sec Ϫ1 , and a relative humidity of 70%). Knock-out Arabidopsis lines containing a T-DNA insert in the Zat12 gene (KO-Zat12; obtained through the SIGnAL project; signal.salk.edu/ tabout.html) were outcrossed and selfed to check for segregation and to obtain pure homozygous lines as described (13,14). Analysis of Zat12 knock-out and segregation was performed by PCR and genomic DNA blots (13). Screening for expression of Zat12 by RNA blots was performed with leaf tissues obtained from wounded and control wild type and knock-out plants 1 h following wounding. Wounding of plants was performed with needles as described (19). Transformation of Arabidopsis plants was performed as described (14,20), and transgenic plants were screened by RNA blots (14). All experiments were performed in triplicate and repeated at least three times (with the exception of DNA arrays, which were performed as described below).
Molecular and Biochemical Analysis-RNA and protein were isolated and analyzed by RNA and protein blots as described previously (13,14). Detection of protein oxidation was performed with the OxyBlot protein oxidation kit (Chemicon International, Temecula, CA) as recommended by the manufacturer. The identity of the major protein bend oxidized in KO-Zat12 plants was determined by immunoprecipitation with an antibody against Rubisco as described (21). RNA staining or a ribosomal 18 S rRNA probe were used to control for RNA loading. Coomassie Blue staining of protein gels was used to control for protein loading. Zat12 (At5g59820), Zat7 (At3g46080), and WRKY25 (At2g30250) detection by RNA blots was performed with gene-specific probes. Full-length clones for WRKY25 were obtained from RIKIN (www.brc.riken.go.jp/lab/epd/Eng/index.html), and full-length cDNA clones for Zat7 and Zat12 were cloned by reverse transcription PCR using mRNA prepared from cells 1 h following wounding (13,19). Clones were sequenced and compared with genomic sequences of Zat12 and Zat7. Gene-specific probes (ϳ200bp) were prepared by PCR according to (22) using the following primers: Zat12 left, 5Ј-CACAAACCA-CAAGAGGATCATTTC-3Ј and Zat12 right, 5Ј-GACGTTTTCACCTTCT-TCATCAAT-3Ј; Zat7 left, 5Ј-TCAAAACCCTAG AAGTCACTA-3Ј and Zat7 right, 5Ј-CAAGAAGTGATGGATTGTC AC-3Ј; and WRKY25 left, 5Ј-AGAAATCTTAAAGTTGTCTCCTTT-3Ј and WRKY25 right, 5Ј-TGG-AAACGTTCCTGTTGTTGGAG-3Ј.
DNA Chip Analysis-In two independent experiments, RNA was isolated from 40 -50 wild type, KO-Zat12, or KO-Apx1 plants (a total of 80 -100 plants per line, in triplicate) grown under controlled conditions as described above. This RNA was used to perform chip analyses (Arabidopsis ATH1 chips; Affymetrix, Santa Clara, CA) at the University of Iowa DNA facility (dna-9.int-med.uiowa.edu/microarrays.htm). Conditions for RNA isolation, labeling, hybridization, and data analysis are described (13,14). Comparative analysis of samples was performed with the GeneChip mining tool version 5.0 and the Silicon Genetics GeneSpring V 5.1. Some of the comparison results were confirmed by RNA blots.
Stress Assays-For the analysis of oxidative stress tolerance of transgenic plants constitutively expressing Zat12, Zat7, and WRKY25, seeds of wild type and transgenic lines were surface-sterilized with bleach and placed in rows on 1% agar plates (0.5ϫ Murashige and Skoog medium) containing different concentrations of paraquat (Sigma). Each row of seeds placed on a plate was divided into two parts, wild type seeds and seeds of transgenic plants. Thus, the different seeds were placed side by side on the same plate. Plates were maintained vertically in a growth room (21-22°C, constant light, and 80 -100 mol m Ϫ2 sec Ϫ1 ), and the percentage of germination and root length were scored 5 days after seed sterilization and plating. Heat shock and light stress were performed as described (13). Osmotic stress and H 2 O 2 and paraquat stresses were performed by subjecting 5-day-old seedlings grown in liquid culture (0.5ϫ Murashige and Skoog medium, 21-22°C, and 100 mol m Ϫ2 sec Ϫ1 ) to polyethylene glycol (PEG 6000, 0 -2%), paraquat (0 -1 M), or H 2 O 2 (0 -20 mM) for 0, 0.5, 1, and 2 h.

RESULTS
Expression Analysis of Zat12, Zat7, and WRKY25-To examine the correlation between Zat12, Zat7, WRKY25, and Apx1 expression in response to oxidative stress or different abiotic stresses, we subjected wild type plants to H 2 O 2 stress, heat shock, a moderate level of light stress (400 mol m Ϫ2 sec Ϫ1 ), wounding, paraquat (a superoxide-generating agent), and osmotic stress. As shown in Fig. 1 application. In contrast, a moderate level of light stress or osmotic stress did not enhance the expression of these transcripts. The level of transcripts encoding Apx1 was elevated by all treatments, suggesting specifically that oxidative stress (H 2 O 2 or paraquat), wounding, and heat shock (Fig. 1) present a clear correlation between Zat12, Zat7, WRKY25, and Apx1 expression. Analysis of the promoter regions of Zat12, Zat7, and WRKY25, performed with the plantCARE software (sphinx.rug.ac.be:8080/PlantCARE/cgi/index.html), revealed that a number of putative DNA binding sites are common among all three genes and Apx1 (Table I). These include a number of light response elements (not shown in Table I), methyl jasmonate (MeJA), ethylene, gibberellin, and salicylic acid response elements, a wound response element (WUN), a MYB-binding site, and the heat shock transcription factor binding site (heat shock element or HSE; Table I). In addition, two conserved motifs with an unknown function (5Ј-GAGACGCG-GTGACAC-3Ј and 5Ј-TCGTCCCAGCC-3Ј) were identified in the promoters of Zat12, Zat7, WRKY25, and Apx1 using the MEME software (meme.sdsc.edu/meme/website/intro.html). Because these motifs were found in the promoters of all four genes, it is possible that they are involved in regulating the expression of these genes during abiotic stress or oxidative stress (Fig. 1). In addition, the expression of Zat12, Zat7, WRKY25, and Apx1 during wounding or heat shock could be explained by the presence of the wound response element (WUN) or heat shock element (HSE) binding sites in the promoters of these genes. Interestingly, the expression of Zat12 was not elevated in response to a moderate level of light stress (400 mol m Ϫ2 sec Ϫ1 ). This result conflicts with a previous report on the expression of Zat12 during high light stress (400 -1800 mol m Ϫ2 sec Ϫ1 ; Ref. 18) and with the presence of many different light response elements in the promoter of Zat12 (light response elements in the promoters of Zat7 and WRKY25 were also found to be non-responsive to the same moderate light stress treatment; Fig. 1).
Time course analysis of Zat12, Zat7, WRKY25, and Apx1 expression during H 2 O 2 stress (Fig. 2) revealed that the steadystate level of transcripts encoding Zat12, Zat7, and WRKY25 is elevated in plants prior to the elevation in Apx1 expression. In addition, the expression of Zat12, Zat7, and WRKY25 was transient and declined within 2 h of H 2 O 2 application. The expression of Zat12 was found to further decline to an undetectable level 4 h following the application of H 2 O 2 (not shown). These findings are in accordance with previous reports on the transient expression of Zat12 during cold stress, wounding, and anoxic stress (16,17,23) and the transient expression of Zat7 during wounding (16).
Analysis of Zat12, Zat7, and WRKY25 in Transgenic Plants-To test the function of Zat12, Zat7, and WRKY25 in plants, we expressed full-length cDNA clones for these putative transcription factors in transgenic plants. For this purpose, we used the 35S-CaMV promoter. We then tested the tolerance of seedlings obtained from transgenic plants that constitutively express Zat12, Zat7, or WRKY25 to oxidative stress using a plate assay that measures root length and percentage of ger-  mination of seedlings in the presence or absence of the superoxide-generating agent paraquat (14). As shown in Fig. 3, seedlings of transgenic plants expressing Zat12 or Zat7 were more tolerant than seedlings of wild type plants to the oxidative stress applied in this assay. In contrast, seedlings of transgenic plants expressing WRKY25 were not more tolerant than seedlings of wild type plants to this treatment. These results were obtained with at least two independent transgenic lines for each of the different putative transcription factors.
Using RNA blots, we tested the expression of Zat12, Zat7, WRKY25, and Apx1 in all transgenic lines grown under controlled growth conditions. As shown in Fig. 4A, constitutive expression of Zat12, Zat7, or WRKY25 in the different transgenic plants did not result in the elevated expression of Apx1 or any of the three putative transcription factors not controlled by the 35S-CaMV promoter. This result suggests that, under the controlled conditions we used to grow plants, the different putative transcription factors expressed in each of the different lines were unable to elevate the expression of each other or the expression of Apx1. A very high level of expression of Zat7, obtained in certain lines at the homozygous state, resulted in a delayed growth and development phenotype (Fig. 4B). However, this level of expression did not elevate the expression of Zat12 or WRKY25 and suppressed the expression level of Apx1 (Fig. 4B). A similar phenotype was not observed in transgenic plants expressing Zat12 or WRKY25 (not shown).

Microarray Analysis of Transgenic Plants Constitutively Ex-
pressing Zat12-Because Zat12 expression enhanced the tolerance of plants to oxidative stress ( Fig. 3) but did not result in the enhanced expression of Zat7, WRKY25, or Apx1 (Fig. 4A), we examined transgenic plants constitutively expressing Zat12 by microarrays to identify transcripts that may be involved in the response of plants to ROS (elevated in Zat12-expressing plants in the absence of an external stimuli). For these studies, we used leaf tissues of 2-week-old plants grown under controlled conditions and compared the pattern of transcript expression between transgenic plants expressing Zat12, wild type plants, and knock-out plants deficient in Apx1 (KO-Apx1). As shown in Table II, ten different transcripts that were elevated in transgenic plants expressing Zat12 were also elevated in KO-Apx1 plants (cutoff 0.8 log 2 -fold; transcripts elevated in Zat12-expressing plants as well as in KO-Apx1 plants are indicated in boldfaced type). It is possible that the expression of these transcripts in Apx1-deficient plants is controlled by Zat12. Our microarray experiments further confirmed that the expression of Zat7, WRKY25, or Apx1 is not elevated in transgenic plants expressing Zat12 (Table II; Fig. 4A). The steadystate level of a number of transcripts with a putative signaling function was elevated in transgenic plants expressing Zat12. These included a monomeric G-protein, MAPK kinase 4, a number of putative transcription factors (i.e. TINY, MYB, and zinc finger proteins), two different kinases, and a calcium- binding protein. Transcripts related to ROS metabolism enhanced in plants constitutively expressing Zat12 included superoxide-generating NADPH oxidase, peroxidase 2a, and glutathione S-transferase. Transcripts related to pathogen response and auxin, ethylene, and methyl jasmonate signaling were also elevated in Zat12-expressing plants (Table II). Supplementary Table I, available in the on-line version of this article, lists all transcripts elevated in transgenic plants expressing Zat12 (cutoff 0.5 log 2 ).
Analysis of Knock-out Plants for Zat12-To test the function of Zat12 in plants during oxidative stress, we obtained and purified Zat12-deficient knock-out lines. When grown under controlled conditions, Zat12-deficient plants were similar in their growth and appearance to wild type plants (not shown). However, when the oxidative stress response of KO-Zat12 plants was compared with that of wild type plants (tested with H 2 O 2 or paraquat), it was found that the steadystate level of transcripts encoding Zat7, WRKY25, and Apx1 was not elevated in KO-Zat12 plants during oxidative stress (Fig. 5A). The expression of Apx1 was, however, elevated in KO-Zat12 plants in response to a moderate level of light stress (Fig. 5A).
To test whether the suppression of Zat7, WRKY25, and Apx1 during oxidative stress (Fig. 5A) resulted in greater damage to cells during oxidative stress, we used a protein blot approach to detect protein oxidation in plants. We first tested protein oxidation in wild type plants during H 2 O 2 stress (10 mM, 1 h), and found that the major protein bend oxidized in Arabidopsis seedlings subjected to this treatment corresponds in its molecular weight to that of the large subunit of Rubisco. Immunoprecipitation assays of Rubisco in protein extracts obtained from treated and untreated plants confirmed that the oxidized protein is indeed Rubisco 2 (not shown; see Ref. 24 for a detailed study of Rubisco protein oxidation during oxidative stress). As shown in Fig. 5B, the H 2 O 2 -induced oxidation of a Rubisco large subunit in KO-Zat12 plants was higher than that in wild type plants subjected to the same H 2 O 2 stress. This finding suggests that the lack of Apx1 expression in knock-out Zat12 plants during oxidative stress results in greater oxidative damage to Rubisco protein, providing further evidence that cytosolic Apx1 is involved in protecting the chloroplast from oxidative stress (9,13). DISCUSSION Enhanced expression of transcripts encoding different regulatory proteins, e.g. 2-component histidine kinase, different receptor-like protein kinases, WRKY transcription factors, calcium-binding proteins, calmodulin-like proteins, and MAPKs, was associated with oxidative stress in plants (11)(12)(13)(14). However, genetic evidence supporting a regulatory role for many of these proteins during oxidative stress was not presented. In Arabidopsis, two different MAP kinases (MAPK3 and MAPK6) were shown to be involved in H 2 O 2 responses, and two different zinc finger proteins (Lsd1 and Lol1) were shown to have an antagonistic effect on cytosolic copper/zinc-superoxide dismutase expression during pathogen response (6,25). In addition, constitutive expression of a heat shock transcription factor (HSF3) in transgenic Arabidopsis plants was shown to enhance the expression of Apx1 and Apx2 in the absence of stress (15). Here, we report that the zinc finger protein Zat12 is required for cytosolic Apx1 expression during oxidative stress (Fig. 5). Furthermore, we show that Zat12 is also essential for the expression of Zat7 and WRKY25 and that these putative transcription factors are involved in the response of plants to oxidative stress (Figs. 3 and 5). The elevation in Zat12, Zat7, and WRKY25 expression in cells prior to the elevation in Apx1 expression during oxidative stress (Fig. 2) and the lack of Zat7, WRKY25, and Apx1 expression during oxidative stress in knock-out Zat12 plants (Fig. 5) provide strong evidence that Zat12, Zat7, and WRKY25 are integral components of the oxidative stress response signal transduction pathway of Arabidopsis. Based on our findings (Fig. 5), we propose that Zat12 acts upstream of Zat7, WRKY25, and Apx1 on the ROS signal transduction pathway of Arabidopsis (Fig. 6).
Interestingly, constitutive expression of Zat12, Zat7, or WRKY25 did not enhance the expression of Apx1 in the absence of stress (Fig. 4A). This result suggests that an additional factor(s), unknown at present, may be required to enable the expression of Apx1 in these plants. This factor may only be present in cells during oxidative stress (Fig. 6). Because Zat12, Zat7, and WRKY25 are transiently induced in cells during stress (Refs.16 -18; Fig. 2), it is possible that their expression is coordinated with that of other factors transiently induced during stress and that the absence of these factors in transgenic plants grown under controlled conditions prevented the induction of Apx1 (see also Ref. 26 for a discussion on expressing inducible transcription factors in plants). To test this possibility, we applied H 2 O 2 stress to wild type plants and transgenic plants that constitutively express Zat12 (similar to the treatment shown in Fig. 2), and compared the expression of Apx1 between wild type plants and transgenic plants that constitutively express Zat12. However, the expression of Apx1 in transgenic plants constitutively expressing Zat12 in response to oxidative stress was only slightly higher than that of wild type plants (1.5-2-fold higher than wild type; not shown). Further studies are therefore required to identify the factors involved in Apx1 expression during stress and determine whether constitutive expression of Zat12, Zat7, and/or WRKY25 in transgenic  (Table II). The enhanced expression of transcripts encoding an NADPH oxidase gene in Zat12-ex-  Table II), was also found to be elevated in plants subjected to cold or salt stress, stresses that enhance the expression of Zat12, as well as the expression of different ROS-scavenging enzymes (30). These findings further support the hypothesis that Zat12 expression can enhance the expression of this NADPH gene during abiotic stress. Compared with seedlings of wild type plants or plants expressing WRKY25, seedlings of plants expressing Zat12 or Zat7 were tolerant to oxidative stress applied on agar plates by paraquat (Fig. 3). In a previous study, seedlings of tobacco plants expressing a constitutively activated form of the oxidative stress signal transduction protein ANP1, a MAPK similar to MAPK3/6 in Arabidopsis, were found to be more tolerant than wild type seedlings to different abiotic stresses such as freezing, heat shock, and salt stress (25). Our findings suggest that additional components of the oxidative stress signal transduction pathway of Arabidopsis could be used in a similar manner to enhance the tolerance of plants to oxidative stress. Because Zat12 expression in transgenic plants did not activate multiple defense mechanisms in plants in the absence of stress (Table II) and did not result in a deleterious side effect on plant growth and yield (not shown), Zat12 may be an ideal signal transduction protein to express in plants and enhance their tolerance to oxidative stress or, potentially, other abiotic stresses. Further studies examining the tolerance of Zat12-and Zat7-expressing plants to different abiotic stresses may reveal whether these proteins could be used for different biotechnological applications such as the enhancement of plant tolerance to biotic or abiotic stress.
In contrast to many of the different transcription factors characterized in plants, the steady-state level of transcripts encoding Zat12 is elevated in Arabidopsis in response to a very large number of different biotic and abiotic stresses. These include stresses such as heat shock, salt, cold, wounding, pathogen, and high light (Refs. 12, 13, and 16 -18, as well as a search of stress response Arabidopsis microarray results available at www.arabidopsis.org/servlets/Search). Common to these stresses, as well as to other stresses that do not enhance Zat12 expression, is the accumulation of ROS in cells during different stages of stress and stress recovery (31). Although it is not known which signals are involved in enhancing Zat12 expression in cells, it is tempting to speculate that a combination of different signals such as ROS and/or different stress response hormones control the expression of Zat12 in cells during stress. Analysis of the Zat12, Zat7, and WRKY25 promoters (Table I) supports a link between different stress hormones and Zat12 expression. However, with the exception of the heat shock factor-binding site that may regulate Zat12 expression during heat shock or oxidative stress, no known DNA binding site for ROS responses was identified in the promoter of Zat12. We are currently using Zat12 promoter-luciferase fusions to study the Zat12 promoter and isolate different mutants deficient in Zat12 expression during stress. Therefore, in the absence of stress, constitutive expression of Zat12 in plants does not enhance the expression of Zat7, WRKY25, or Apx1 (Fig. 4).