A regulatory cascade involving class II ETHYLENE RESPONSE FACTOR transcriptional repressors operates in the progression of leaf senescence.

This work demonstrates the proteasome-mediated regulation of class II ERF transcriptional repressors and involvement of these factors in the progression of leaf senescence. ABSTRACT Leaf senescence is the final process of leaf development that involves the mobilization of nutrients from old leaves to newly growing tissues. Despite the identification of several transcription factors involved in the regulation of this process, the mechanisms underlying the progression of leaf senescence are largely unknown. Herein, we describe the proteasome-mediated regulation of class II ETHYLENE RESPONSE FACTOR (ERF) transcriptional repressors and involvement of these factors in the progression of leaf senescence in Arabidopsis thaliana (Arabidopsis) . Based on previous results showing that the Nicotiana tabacum ERF3 (NtERF3) specifically interacts with a ubiquitin-conjugating enzyme, we examined the stability of NtERF3 in vitro and confirmed its rapid degradation by plant protein extracts. Furthermore, NtERF3 accumulated in plants treated with a proteasome inhibitor. The Arabidopsis class II ERFs AtERF4 and AtERF8 were also regulated by the proteasome and increased with plant aging. Transgenic Arabidopsis plants with enhanced expression of NtERF3, AtERF4, or AtERF8 showed precocious leaf senescence. Our gene expression and chromatin immunoprecipitation analyses suggest that AtERF4 and AtERF8 targeted EPITHIOSPECIFIER PROTEIN/EPITHIOSPECIFYING SENESCENCE REGULATOR gene and regulated the expression of many genes involved in the progression of leaf senescence . By contrast, an aterf4 aterf8 double mutant exhibited delayed leaf senescence. Our results provide insight into the important role of class II ERFs in the progression of leaf senescence. (Bio-Rad) with a CFX96 real-time PCR system (Bio-Rad). Transcript levels were detected in triplicate using a standard curve derived from the reference sample. The relative values of the transcripts were normalized to the UBQ1 level.


INTRODUCTION
Leaf senescence, which is characterized by progressive yellowing, is the final stage of leaf development and involves the mobilization of nutrients from old leaves to newly growing tissues. The progression of leaf senescence requires programmed cell death combined with the cessation of photosynthesis, organelle breakdown, and protein degradation. The regulation of leaf senescence depends largely on the developmental age of plants, although it is also influenced by various external stimuli (Gan and Amasino, 1997). The detection of internal and external signals activates various processes mediated by signaling molecules such as plant hormones and reactive oxygen species (ROS; Buchanan-Wollaston et al., 2003;Lim et al., 2007).
The progression of leaf senescence is associated with the downregulation of genes involved in chlorophyll biosynthesis, carbon metabolism, and photosynthesis and the upregulation of genes involved in responses to hormones, ROS, and various stresses (Gepstein et al., 2004;Lin and Wu, 2004;Buchanan-Wollaston et al., 2005;van der Graaff et al., 2006;Balazadeh et al., 2008;Breeze et al., 2011). The coordinated regulation of gene expression during leaf senescence depends on the combined action of several families of transcription factors (TFs). In particular, genes coding for NAM/ATAF/CUC2 (NAC), zinc finger, WRKY, MYB, and APETALA2/ETHYLENE RESPONSE FACTOR (AP2/ERF) TFs are transcriptionally upregulated during leaf senescence (Lin and Wu, 2004;Buchanan-Wollaston et al., 2005;Breeze et al., 2011).
AP2/ERF TFs consist of 146 members in Arabidopsis and are involved in responses to various external and internal stimuli. Several AP2/ERFs modulate responses to leaf senescence-associated signaling molecules such as ROS, ethylene, jasmonic acid (JA), abscisic acid (ABA), and cytokinin (Nakano et al., 2006;Mizoi et al., 2012). Gain-of-function studies suggest that RAV1 and C-REPEAT BINDING FACTOR (CBF)/DEHYDRATION RESPONSIBLE ELEMENT BINDING1 (DREB1) positively and negatively regulate leaf senescence, respectively (Sharabi-Schwager et al., 2010;Woo et al., 2010). However, knowledge of the molecular mechanisms of these AP2/ERFs in the regulation of leaf senescence is lacking.
Class II ERFs are characterized by the ERF-associated amphiphilic repression (EAR) motif (Ohta et al., 2001). Because class II ERFs repress target gene transcription in the presence of ERF activators in transient gene expression assays (Fujimoto et al., 2000;Ohta et al., 2000Ohta et al., , 2001, understanding the regulation of class II ERFs is key to elucidating the complex mechanisms underlying AP2/ERF-mediated gene regulation. In addition to studying the stress response-associated transcriptional control of class II ERF genes, (Suzuki et al., 1998;Yamamoto et al., 1999;Fujimoto et al., 2000;Kitajima et al., 2000;Nishiuchi et al., 2003), we previously analyzed the post-translational regulation of class II ERFs and showed that a class II ERF from Nicotiana tabacum, NtERF3, physically interacts with the ubiquitin-conjugating (UBC) enzyme NtUBC2 (Koyama et al., 2003). Coexpression of the dominant-negative form of NtUBC2 enhances the repressive activity of NtERF3 in transient assays, suggesting that NtUBC2 may regulate NtERF3 activity. By contrast, NtUBC2 does not interact with NtERF2, a transcriptional activator. The function of NtUBC2 in the ubiquitin-proteasome system (Dreher and Callis, 2007;Vierstra, 2009) suggests that NtERF3 may be regulated by proteolysis.
In the present study, we demonstrate the involvement of proteasomes in the control of NtERF3 and two homologous Arabidopsis class II ERFs, AtERF4 and AtERF8. We also performed functional and gene expression analyses to clarify the role of these class II ERFs in the progression of leaf senescence. resin (Supplemental Fig. S1). Each protein was incubated for 0, 15, 30, and 60 minutes in a cultured tobacco XD6S cell extract, and protein stability was analyzed with immunoblotting. Whereas NtERF3 was extremely unstable in the tobacco cell extract solution (Fig. 1A), NtERF2 and GFP were stable for 60 minutes (Figs. 1B and 1C).
NtERF3 levels remained stable in a bovine serum albumin solution (Fig. 1D), suggesting that NtERF3 is specifically degraded by plant proteins.
To identify the amino acid residues involved in the stability of NtERF3, we produced truncated NtERF3 proteins in E. coli ( Fig. 1E and Supplemental Fig. S1). NtERF3 amino acid residues 83 to 95 contain a PEST motif, which is involved in the instability of proteins (Asher et al., 2006); however, deletion of the PEST motif did not affect the stability of NtERF3 in vitro (Fig. 1E).

Low accumulation of NtERF3 in plant cells
To examine the stability of NtERF3 in plants, we individually fused the coding sequences (CDS) for the respective regions of NtERF3 with the nuclear localization signal (NLS)-GFP sequence under the control of the cauliflower mosaic virus 35S promoter (Pro35S:NLS-GFP), and the resulting fusion genes were transformed into cultured tobacco XD6S cells (Supplemental Fig. S2A). More than 100 kanamycin-resistant tobacco calluses for each construct were examined for GFP florescence. We detected no GFP signal in the Pro35S:NLS-GFP-NtERF3 and These results suggest that NtERF3 levels were low in plant cells and that the NtERF3 (83/190) region may determine protein instability.

Proteasomal degradation of class II ERF repressors
To characterize the mechanism underlying the proteolytic degradation of NtERF3, we treated Pro35S:NLS-GFP-NtERF3 Arabidopsis plants with MG132, a proteasome inhibitor. Fluorescence microscopy and immunoblot analyses showed that MG132 treatment enhanced the accumulation of NLS-GFP-NtERF3 (Fig. 3A).
We further investigated the role of proteasomes in the stability of the NtERF3 homologues AtERF4 and AtERF8 (Ohta et al., 2001). Hemaglutinin (HA)-tagged Immunoblot analysis showed that MG132 treatment increased AtERF4-HA and AtERF8-HA levels, whereas dimethyl sulfoxide (DMSO) treatment had no effect (Figs. 3B and 3C). Furthermore, an upper-shifted band corresponding to AtERF4-HA suggested post-translational modification of AtERF4 (Fig. 3C). These results provided evidence of the involvement of proteasome in the control of NtERF3, AtERF4, and AtERF8 stabilities.
Interestingly, AtERF4-HA and AtERF8-HA accumulated in aging plants in the absence of MG132 treatment (Fig. 3D). However, unlike MG132-treated plants, no shift of AtERF4-HA was detected in older plants. These results indicated that aging also stabilized AtERF4 and AtERF8.

Cell death and precocious leaf senescence induced by class II ERFs.
Our study of the stability of ERF indicated that ectopic expression of class II ERF repressor genes induced a phenotype characterized by cell death and precocious leaf senescence (Fig 4). We therefore attempted to characterize the functions of class II

Regulation of genes involved in leaf senescence by AtERF4.
Because ectopic expression of the class II ERF genes examined had a similar effect on inducing cell death and precocious leaf senescence, these ERFs might regulate expression of a common set of downstream genes. Previous studies revealed the induction of AtERF4 expression by ethylene, ABA, and JA, which regulate the progression of leaf senescence (Fujimoto et al. 2000;McGrath et al., 2005;Yang et al., 2005), and then AtERF4 was used as a model for further characterization of downstream genes. Microarray analysis detected 929 genes transcriptionally increased more than 2-fold (P value of dependent t-test < 0.05; false discovery rate [FDR] < 0.04225; see Materials and Methods) and 687 genes decreased less than half (P < 0.05) We searched for TF binding sites in the 1,000-bp region upstream of the transcription initiation site of the genes upregulated in Pro35S:AtERF4-HA plants and found that the group of genes containing a W-box, which is a binding motif of WRKY TFs, was overrepresented (Supplemental Table S1). Consistently, 17 of 71 WRKY genes were significantly upregulated in Pro35S:AtERF4-HA plants (Supplemental   Tables S2 and S3). Among them, WRKY30, WRKY53, and WRKY75 were positive regulators of leaf senescence, whereas WRKY18, WRKY40, and WRKY60 were involved in the basal defense response against pathogen attack (Xu et al., 2006;Miao and Zentgraf, 2007;Besseau et al., 2012;Li et al., 2012). In our RT-PCR analysis, WRKY18, WRKY30, WRKY40, WRKY53, WRKY60, and WRKY75 transcripts were increased in Pro35S:AtERF4-HA and Pro35S:AtERF8-HA plants (Fig. 6), indicating that AtERF4 and AtERF8 regulated the expression of these genes. In the leaves of the aterf4 aterf8 mutant, the expression of genes involved in the positive regulation of leaf senescence was downregulated, whereas that of the negative regulator genes was increased (Fig. 8). WRKY30, WRKY53, and WRKY75 transcripts were low in the mature leaves of the wild type and subsequently increased in older leaves but were maintained at low levels in mature and old leaves of aterf4 aterf8 mutants. Conversely, IAA3/SHY2, IAA6/SHY6, and ESP/ESR transcripts in mature leaves of the aterf4 aterf8 mutant were increased compared with those of the wild type, whereas these transcripts were gradually reduced in older leaves (see Fig. 8 and Supplemental Fig. S5). Taken together, these results suggest that the delayed leaf senescence phenotype of the aterf4 aterf8 mutant was responsible for changes in the expression profile.
Delayed leaf senescence in aterf4 aterf8 plants was unlikely to have resulted from general growth retardation because flowering time in this mutant was comparable to that of wild-type plants (Supplemental Fig. S6). Furthermore, no defects in the general components of the ethylene and JA response pathways in the aterf4 aterf8 mutant was detected-it progressed at a pace similar to that of wild type in the presence of 1-aminocyclopropanecarboxylic acid, a precursor of ethylene, and methyl jasmonate (Supplemental Fig. S7).

Binding of AtERF4 and AtERF8 to ESP/ESR
Because AtERF4 and AtERF8 function primarily as transcriptional repressors (Ohta et al., 2001), we presumed that their target genes might be downregulated in In addition, the ChIP assay using Pro35S:NLS-GFP-HA Arabidopsis plants did not detect enrichment of the ESP/ESR sequences (see Fig. 9). These results suggested specific binding of AtERF4 and AtERF8 to ESP/ESR in plant cells.

Induction of AtERF4 and AtERF8 expression during leaf senescence
We and AtERF8 transcripts began 5 weeks after germination and preceded leaf yellowing, but this increase was delayed in the quintuple tcp mutant (see Fig. 10C). These results suggested that AtERF4 and AtERF8 expression was induced during leaf senescence.    (Rushton et al., 2010). WRKY53 and WRKY75 are involved in responses to biotic stress and nutrient starvation, respectively (Devaiah et al., 2007;Miao and Zentgraf, 2007). WRKY18, WRKY40, and WRKY60 enhance the basal defense system against pathogens (Xu et al., 2006). In addition, our microarray analysis suggested that which regulate responses to ethylene, JA, ABA, hydrogen peroxide, and dehydration (Solano et al., 1998;Sakamoto et al., 2004;Davletova et al., 2005;Chini et al., 2007;Ciftci-Yilmaz et al., 2007;Thines et al., 2007;Jung et al., 2008;Kim et al., 2009;Guo and Gan, 2011). Because these stresses affect the progression of leaf senescence, we postulate that these class II ERFs may integrate aging and stress signaling pathways. We demonstrate that aging stimulates expression of AtERF4 and AtERF8 in leaves and increases ERF accumulation. In addition to elucidating the stress response-associated transcriptional control of class II ERFs (Suzuki et al., 1998;Yamamoto et al., 1999;Fujimoto et al., 2000;Kitajima et al., 2000;Nishiuchi et al., 2003), our study adds some insights to the developmental regulation of these ERFs. Whereas selective protein turnover is regulated by proteasomes during leaf senescence, bulk protein degradation is processed by autophagy (Vierstra, 2009). Prior studies have described the role of several components of the ubiquitin-proteasome system (Woo et al., 2001;Gepstein et al., 2004;Lin and Wu, 2004;Buchanan-Wollaston et al., 2005;van der Graaff et al., 2006;Peng et al., 2007, Raab et al., 2009Miao and Zentgraf, 2010;Breeze et al., 2011). In particular, WRKY53 is  In summary, we showed the proteasomal regulation, age-dependent accumulation, and regulatory cascade of class II ERFs during leaf development. Our results provide cues of a sophisticated mechanism for the progression of leaf senescence involving class II ERFs.

Plant materials and growth conditions
Growth conditions and the transformation of tobacco XD6S cells and Arabidopsis plants have been described previously (Yamamoto et al., 1999;Koyama et al., 2010).

In vitro degradation assay
The 6xHis-tagged NtERFs and GFP were produced in E. coli, purified using a nickel-chelating resin according to manufacturer instructions (Amersham Pharmacia), and dialyzed against a solution containing 20 mM Tris (pH 8.0) and 100 mM NaCl.
The cell-free degradation assay was modified from previous reports as described below (Osterlund et al., 2000). For preparation of the extract solution, cultured tobacco XD6S cells were ground in liquid nitrogen, suspended in a buffer containing 50 mM Tris (pH 7.5), 10 mM NaCl, 10 mM MgCl 2 , 5 mM dithiothreitol, 2 mM ATP, and COMPLETE ethylenediaminetetraacetic acid (EDTA)-free protease inhibitor cocktail (Promega) and centrifuged at 15,000 rpm for 10 min at 4°C. The supernatant was diluted to 4 mg/mL to make the cell extract solution. Aliquots containing 100 ng recombinant NtERFs and GFP were added into 200 µL cell extract solution, incubated at room temperature for 0, 15, 30, or 60 minutes, mixed in an equal volume of 2X sodium dodecyl sulfate (SDS) loading buffer (Tris pH 6.8, 2% SDS, 1 mM EDTA, 2-mercaptoethanol), and boiled to terminate the reaction. The proteins were separated by SDS-polyacrylamide gel electrophoresis (PAGE), blotted onto Immobilon-P membranes (MILLIPORE), and subjected to immunoblot analysis.

Preparation of tobacco and Arabidopsis proteins
For preparation of plant protein samples, tobacco XD6S callus lines at 2 weeks after the inoculation on new plates and Arabidopsis seedlings grown on Murashige-Skoog plates were ground in liquid nitrogen, suspended in extraction buffer (Tris pH 6.8, 8M urea, 0.5% SDS, 1 mM EDTA, 2-mercaptoethanol), incubated for 2 minutes at 95°C, and mixed with one-third volume of SDS/urea loading buffer (Tris pH 6.8, 8M urea, 2% SDS, 1 mM EDTA, 2-mercaptoethanol). Aliquots containing 10 µg and 30 µg of protein from tobacco XD6S cells and Arabidopsis plants, respectively, were separated using SDS-PAGE, blotted onto Immobilon-P membranes (MILLIPORE), and subjected to immunoblot analysis.

Antibodies and immunoblot analysis
For production of anti-NtERF2 and anti-NtERF3 antibodies, the recombinant NtERF2 and NtERF3 proteins fused with maltose-binding protein were produced in E. Immunoreactive proteins were detected using the ECL plus western blotting kit (GE Healthcare).

Microscopy
The fluorescence images in Figs. 2A and 2C were, respectively, monitored with a BHS-RFC (Olympus) as described previously (Ohta et al., 2000) and a BZ-9000 (KEYENCE) with a GFP-Bandpass filter at a fixed one-third-second exposure. Bright field images were obtained with an MZ FL III (Leica).

Gene expression analysis
Total RNA was prepared from tobacco XD6S cells and Arabidopsis plants using Microarray analysis was performed with the two-color method using aliquots of total RNA from 4 biological replicates of 2-week-old Pro35S:AtERF4-HA (line #20) and Pro35S:NLS-GFP-HA Arabidopsis plants with an Agilent Arabidopsis V3 (4x44k) microarray. We used previously described preparation and statistical analyses (Koyama et al., 2010) with the following exceptions. Only genes with average detection values of ≥1.5 in both "test" and "reference" samples were analyzed. The P value for each gene was calculated using a dependent t-test. To estimate the FDR, we calculated the Q value from the P value using QVALUE software with the default settings (Storey and Tibshirani, 2003) and selected up-/downregulated (>2-fold/<0.5-fold) genes with a P value of ≤0.05 (FDR < 0.04225). In Fig. 5, the transcript level of each gene at different ages was calculated as relative to that at 19 days after sowing (Breeze et al., 2011), and a clustered heat map was prepared using Cluster3 software (Eisen et al., 1998). To evaluate over-representation of some gene lists among up-/downregulated genes, we performed a binomial test using R (http://www.r-project.org/).

ChIP
Nuclear extract was prepared from 3-week-old Arabidopsis plants, subjected to sonication, and immunoprecipitated with 500 ng anti-HA antibody (clone 3F10; Roche) as described previously (Koyama et al., 2010). The chromatin precipitated was reverse-cross-linked and purified with ethanol and served as a template for PCR (CFX96 real-time PCR system; BioRad) using an appropriate set of primers (Supplemental Table S5). The values were calculated with a standard curve generated from the input sample and normalized using a background value determined from the eIF4A sequence. Figure 10 Table S1. cis elements enriched in the genes regulated in 35S:AtERF4-HA plants.          Transcript levels were determined with RT-PCR using aliquots of total RNAs from the sixth leaves of wild-type (black squares) and aterf4 aterf8 (white circles) plants at the indicated ages. The values of 4-week-old wild-type leaves were set at 1.
Error bars indicate standard deviation of technical triplicates.   The proteins indicated were produced as fusion proteins with thioredoxin and a 6xHIS tag in E. coli and purified with a nickel affinity column.