Identification of a novel E3 ubiquitin ligase that is required for suppression of premature senescence in Arabidopsis

Authors

  • Sabine Raab,

    1. Molekulare Pflanzenphysiologie, Friedrich-Alexander-Universität Erlangen-Nürnberg, Staudtstrasse 5, D-91058 Erlangen, Germany
    Search for more papers by this author
    • These authors contributed equally to this paper.

  • Gabriele Drechsel,

    1. Molekulare Pflanzenphysiologie, Friedrich-Alexander-Universität Erlangen-Nürnberg, Staudtstrasse 5, D-91058 Erlangen, Germany
    Search for more papers by this author
    • These authors contributed equally to this paper.

  • Maryam Zarepour,

    1. Institut für Pflanzenbiologie, Technische Universität Braunschweig, Spielmannstrasse 7, D-38106 Braunschweig, Germany
    Search for more papers by this author
  • Wolfram Hartung,

    1. Julius-von-Sachs-Institut für Biowissenschaften, Universität Würzburg, Julius-von-Sachs-Platz 2, D-97082 Würzburg, Germany
    Search for more papers by this author
  • Tomokazu Koshiba,

    1. Department of Biological Sciences, Tokyo Metropolitan University, Hachioji, Tokyo 192-0397, Japan
    Search for more papers by this author
  • Florian Bittner,

    1. Institut für Pflanzenbiologie, Technische Universität Braunschweig, Spielmannstrasse 7, D-38106 Braunschweig, Germany
    Search for more papers by this author
  • Stefan Hoth

    Corresponding author
    1. Molekulare Pflanzenphysiologie, Friedrich-Alexander-Universität Erlangen-Nürnberg, Staudtstrasse 5, D-91058 Erlangen, Germany
      (fax +49 9131 852751; e-mail shoth@biologie.uni-erlangen.de).
    Search for more papers by this author

(fax +49 9131 852751; e-mail shoth@biologie.uni-erlangen.de).

Summary

During leaf senescence, resources are recycled by redistribution to younger leaves and reproductive organs. Candidate pathways for the regulation of onset and progression of leaf senescence include ubiquitin-dependent turnover of key proteins. Here, we identified a novel plant U-box E3 ubiquitin ligase that prevents premature senescence in Arabidopsis plants, and named it SENESCENCE-ASSOCIATED E3 UBIQUITIN LIGASE 1 (SAUL1). Using in vitro ubiquitination assays, we show that SAUL1 has E3 ubiquitin ligase activity. We isolated two alleles of saul1 mutants that show premature senescence under low light conditions. The visible yellowing of leaves is accompanied by reduced chlorophyll content, decreased photochemical efficiency of photosystem II and increased expression of senescence genes. In addition, saul1 mutants exhibit enhanced abscisic acid (ABA) biosynthesis. We show that application of ABA to Arabidopsis is sufficient to trigger leaf senescence, and that this response is abolished in the ABA-insensitive mutants abi1-1 and abi2-1, but enhanced in the ABA-hypersensitive mutant era1-3. We found that increased ABA levels coincide with enhanced activity of Arabidopsis aldehyde oxidase 3 (AAO3) and accumulation of AAO3 protein in saul1 mutants. Using label transfer experiments, we showed that interactions between SAUL1 and AAO3 occur. This suggests that SAUL1 participates in targeting AAO3 for ubiquitin-dependent degradation via the 26S proteasome to prevent premature senescence.

Introduction

Leaf senescence gives rise to impressive fall (autumn) foliage in deciduous trees and is necessary for the death of leaves in annual plants. Senescence is developmentally and genetically well-defined, guaranteeing recycling of resources from senescing leaves into young leaves or seeds to maximize growth and reproductive capacity. Leaf senescence represents a process at the organ level, and the yellowing of leaves is accompanied by changes in cellular metabolism and deterioration of cellular structures (Nooden, 1988; Gan and Amasino, 1997; Buchanan-Wollaston et al., 2003; Thomas et al., 2003; Yoshida, 2003; Lim et al., 2007). The observed changes are mediated by viable cells and tissues, and the fact that re-greening can be induced in senescing leaves (in particular by cytokinin) indicates the potential reversibility of leaf senescence (Zavaleta-Mancera et al., 1999a,b).

Onset of senescence is determined by both developmental stage and various environmental factors such as shading, unfavorable temperatures, drought periods and nutritional deprivation. Components of the regulatory network controlling leaf senescence in Arabidopsis have been identified by genetic screens for senescence mutants and by gene expression screens for senescence-associated genes (Yoshida, 2003; Lim et al., 2007). Among these components are the transcriptional regulators WRKY6, WRKY53, NAP and ARF2, the protein kinase EDR1, and the cytokinin receptor AHK3 (Robatzek and Somssich, 2002; Miao et al., 2004; Ellis et al., 2005; Tang et al., 2005; Guo and Gan, 2006; Kim et al., 2006). Cytokinin levels decrease during leaf senescence, and it has been shown that external application of cytokinins or enhanced production of cytokinins in plants by expression of an isopentenyl transferase gene from the SAG12 promoter delays senescence (Gan and Amasino, 1995; McCabe et al., 2001). In contrast to the low cytokinin levels, the levels of ethylene and abscisic acid (ABA) increase during leaf senescence, and these phytohormones are positive regulators of senescence (Gepstein and Thimann, 1980; Yoshida, 2003; Lim et al., 2007). Genetic evidence for the function of ABA signaling components in leaf senescence is scant.

The cross-talk between stress and hormone signaling during the progression of leaf senescence requires tight regulation of signaling components, potentially by protein turnover. Protein degradation is mediated through the ubiquitin–26S proteasome pathway (Sullivan et al., 2003; Vierstra, 2003). Expression of genes encoding polyubiquitin or enzymes of the ubiquitin pathway is up-regulated, ubiquitin conjugation activity is increased, and proteasome activity is maintained during senescence (Park et al., 1998; Yoshida, 2003; Lin and Wu, 2004). Two senescence mutants that are defective in components of the ubiquitin pathway have been described in Arabidopsis. The delayed senescence 1 mutant is defective in the R-transferase AtATE1 that is a component of the N-end rule pathway. The oresara 9 mutant is defective in the F-box protein ORE9 that may be part of an SCF-type E3 complex ligating ubiquitin to target proteins (Woo et al., 2001; Yoshida et al., 2002). Both mutants showed delayed senescence, suggesting that negative regulators of leaf senescence may be targeted by these components for degradation.

Here we show that a mutation in the SAUL1 gene leads to premature senescence in Arabidopsis. We demonstrate that saul1 mutants grown under low light conditions exhibit yellowing of leaves, decreasing chlorophyll content, and enhanced expression of senescence genes. At low but not high photon flux densities, ABA levels were strongly increased. This increase was accompanied by enhanced activity and accumulated protein levels of aldehyde oxidase AAO3, which converts abscisic aldehyde to ABA. We demonstrate ubiquitin ligase activity for SAUL1 and interactions between SAUL1 and AAO3, and suggest that SAUL1 represents a regulatory element that prevents senescence from occurring prematurely by targeting AAO3 for degradation.

Results

Isolation and characterization of saul1 mutant plants

Previously, we described the identification of Arabidopsis genes encoding proteins involved in regulated proteolysis that were transcriptionally regulated by ABA (Hoth et al., 2002). Four of these encode plant U-box proteins (PUB), namely At1g10560 (AtPUB18), At1g20780 (SAUL1/AtPUB44), At1g60190 (AtPUB19) and At3g07360 (AtPUB9), carrying ARM repeats that are important for protein–protein interactions. Among these four PUB-ARM proteins, SAUL1 is structurally unique, because it lacks the characteristic U-box N-terminal domain (UND) that is present in AtPUB18 and AtPUB19, and contains a high number of ARM repeats (Mudgil et al., 2004). We isolated two insertion mutants from the SALK collection corresponding to stock numbers SALK_063974 and SALK_076799 (Alonso et al., 2003) to study the function of SAUL1 in plant development. From segregating T3 lines, we established plants homozygous for the T-DNA insertion identified by PCR genotyping and renamed these saul1-1 and saul1-2. Sequence analysis of the T-DNA flanking regions indicated that the T-DNA is located 2458 bp after the start ATG in exon 3 of SAUL1 in saul1-1, and 630 bp after the start ATG in exon 2 of SAUL1 in saul1-2 mutants (Figure 1a). RT-PCR analyses failed to detect full-length SAUL1 transcripts in saul1 mutants (Figure 1b). saul1-1 mutants contained no additional T-DNA insertion based on the results of TAIL-PCR experiments, but the insertion in the saul1-2 mutants also deleted three additional genes, namely At1g20790, At1g20795 and At1g20800.

Figure 1.

 Senescence phenotype of saul1-1 mutants.
(a) Schematic illustration of the SAUL1 gene structure with the positions of the T-DNA insertions for the two alleles saul1-1 and saul1-2. Exons are shown as black boxes and introns as black lines.
(b) RT-PCR analysis of SAUL1 transcript levels in wild-type and saul1 plants. ACTIN2 transcript levels served as a loading control.
(c) Symptoms of senescence were present in saul1-1 plants that were homozygous for the insertion in SAUL1 (saul1/saul1), but not in plants that were heterozygous for the insertion (SAUL1/saul1) or in wild-type plants (SAUL1/SAUL1). Plants were grown under short-day conditions at a PFD of 60 μmol m−2 sec−1.

When grown on soil, homozygous saul1-1 and saul1-2 mutants exhibited symptoms of early senescence compared to wild-type and heterozygous mutants. Both saul1 mutants showed yellowing of leaves (Figures 1c and S1a), and were identical with respect to development (Figure S1b). To rule out that the phenotype is related to the above-mentioned deletion of additional genes in saul1-2 mutants, we isolated plants that were homozygous for insertions in the respective genes from segregating T3 generations of T-DNA lines corresponding to SALK_133945, SALK_043102 and SALK_101715 (Alonso et al., 2003). Sequence analysis of the T-DNA flanking regions indicated that the T-DNA is located at +918, −52 and +575 bp relative to the start ATG, respectively. The development of these insertion lines was identical to that of wild-type plants (Figure S1c). Transcripts of these three genes could not be detected in wild-type plants, and no ESTs were associated with these genes. Apparently, it is not the deletion of these genes but rather the insertion in SAUL1 that is relevant for senescence in saul1-2 mutants, which is thus an independent allele to saul1-1.

To investigate whether the senescence phenotype of saul1 mutants depends on the photosynthetically active photon flux density (PFD), we grew wild-type and mutant plants on agar. At low PFD, saul1 mutants developed normally until emergence of leaves 3 and 4, but exhibited yellowing of leaves afterwards (Figures 2a and 5h and S1d). The plants did not survive, and this process could not be reversed by high PFD. When plants were grown at higher PFD, the saul1 mutants stayed green without any visible senescence (Figures 2b and 3a,c). Leaf senescence also occurred if saul1-1 mutants were first grown at high PFD and subsequently transferred to low PFD or darkness (Figures 3c and S1g). Of the offspring of heterozygous plants, approximately 25% showed the phenotype, indicating that the mutation is recessive. We confirmed that the light-dependent mutant phenotype segregated with the T-DNA insertion in SAUL1 by back-crossing saul1-1 mutants to wild-type (Figure S1e). At a PFD of 200 μmol m−2 sec−1, we were able to grow saul1 mutant plants to the seed stage to obtain viable homozygous seeds (Figure 2c). However, the saul1 mutants showed reduced growth and commencement of leaf senescence (Figure 2c). To suppress the reduced growth phenotype at that developmental stage, mutant plants had to be grown at a PFD of at least 400 μmol m−2 sec−1 (Figure 2d). To additionally prove that all observed phenotypic changes are due to the mutation in SAUL1, we attempted to complement the phenotypes. For this purpose, we transformed saul1-1 mutants with a CaMV 35S promoter:SAUL1 cDNA construct. Four of the generated saul1-1 35S-SAUL1 plant lines, which showed up to 17-fold induced expression of SAUL1, were subsequently analyzed (Figure 2e). Those plants survived under light conditions that resulted in early senescence of saul1-1 mutants (Figure 2f,g). These plants also grew to the same height as wild-type plants, confirming that all observed saul1 phenotypes were complemented and that the mutation in SAUL1 was the cause of all phenotypic defects.

Figure 2.

 Light dependence of senescence in saul1-1 mutants and complementation of phenotypes.
(a) Offspring of saul1-1 mutants that were homozygous (saul1/saul1) or heterozygous (SAUL1/saul1) for the insertion in SAUL1, and of wild-type plants (SAUL1/SAUL1). Seedlings were grown on agar under long-day conditions at a PFD of 15 μmol m−2 sec−1, and were scored for senescence phenotype after 10 days.
(b) The same experiment as in (a) performed at a PFD of 90 μmol m−2 sec−1.
(c) Wild-type and saul1-1 plants were grown on soil under long-day conditions at a PFD of 200 μmol m−2 sec−1 to monitor the development of plants under higher light conditions.
(d) The same experiment as in (c) performed at a PFD > 400 μmol m−2 sec−1.
(e) Relative expression of SAUL1 in wild-type, saul1-1 and complemented saul1-1/35S-SAUL1 plants was determined by real-time RT-PCR. Data represent mean values ± SE of three replicates.
(f) Complementation of the mutant phenotype at the young seedling stage by transformation of SAUL1 under the control of the CaMV 35S promoter into saul1-1 mutants. Plants were grown under long-day conditions at a PFD of 20 μmol m−2 sec−1, which did not allow survival of saul1-1 mutants.
(g) Complementation of mutant phenotype in adult plants that were grown under long-day conditions at a PFD of 50 μmol m−2 sec−1, which did not allow survival of saul1-1 mutants (top left).

Figure 5.

 Activation and stabilization of AAO3 in saul1 mutants.
(a) Real-time RT-PCR analyses of AAO3 and NCED3 expression in wild-type and saul1 plants at high and low PFDs. The same plant material as in Figure 3(c) was analyzed. ACTIN2 was used as a reference for equal amounts of cDNA. The data for real-time RT-PCR represent mean values ± SD (n = 3).
(b) Aldehyde oxidase activities in saul1 and wild-type plants grown under conditions as in (a). Crude protein extracts (50 μg) were subjected to native PAGE, and aldehyde oxidase activity was visualized in situ using indole-3-carboxaldehyde and 1-naphthaldehyde as substrates.
(c) Western blot analysis of AAO3 protein levels in saul1 mutant and wild-type plants grown under conditions as in (a). Crude protein extracts were subjected to native PAGE followed by immunoblot analysis using anti-AAO3 antibodies.
(d) Aldehyde oxidase activities in saul1 mutant and wild-type plants grown at PFDs of 500 and 250 μmol m−2 sec−1 were detected as in (b).
(e) Aldehyde oxidase activities in wild-type, saul1-1 mutant and saul1-1/35S-SAUL1 plants grown at a PFD of 20 μmol m−2 sec−1 were detected as in (b).
(f) Real-time RT-PCR analysis of relative AAO3 transcript levels in senescing (sl) and non-senescing (nsl) leaves of wild-type plants grown at a PFD of 85 μmol m−2 sec−1 for several weeks. Data are mean values ± SD of three independent experiments.
(g) Detection of aldehyde oxidase activity (top) and Western blot analysis of AAO3 protein levels (bottom) in senescing (sl) and non-senescing (nsl) wild-type leaves from (f).
(h) Western blot analysis of AAO3 abundance during seedling development using anti-AAO3 antibodies. Wild-type and saul1-1 plants were grown at a PFD of 30 μmol m−2 sec−1. The phenotype and AAO3 protein levels in wild-type and saul1-1 plants were monitored at the indicated times.

Figure 3.

 Molecular senescence-related phenotypes of saul1 mutants.
(a) Chlorophyll contents are shown for wild-type (white bars) and saul1-1 plants (black bars) that were grown on agar plates under long-day conditions at the indicated PFDs for 14 days. Photographs of the respective wild-type and saul1 mutant plants are shown on the top of each bar. Data represent mean values ± SD from three independent experiments.
(b) Photochemical efficiencies given as the Fv/Fm ratio for wild-type (white bars) and saul1 mutant plants (black bars) that were grown for 14 days under long-day conditions at the indicated PFDs. Data represent mean values ± SD from three independent experiments.
(c) RT-PCR analyses of chlorophyll biosynthesis, chlorophyll degradation, senescence-associated autophagy and regulatory genes in wild-type and saul1 plants grown under long-day conditions at a PFD of 60 μmol m−2 sec−1 for 10 days and then transferred for an additional 7 days to high PFD (110 μmol m−2 sec−1) or low PFD (20 μmol m−2 sec−1). ACTIN2 was used as a reference for equal amounts of cDNA loaded.

saul1 mutants show early senescence symptoms at the level of chlorophyll content, photochemical efficiency of photosystem II, and gene expression

Leaf longevity of saul1 mutants was assessed by measuring typical physiological markers such as chlorophyll content and the photochemical efficiency of photosystem II as well as molecular markers such as gene expression changes. The PFD dependence of chlorophyll content was determined at day 14 after sowing. At PFDs between 90 and 140 μmol m−2 sec−1, the chlorophyll content did not differ between saul1 and wild-type plants. At a PFD < 70 μmol m−2 sec−1, however, the chlorophyll content started to decrease in saul1 but not in wild-type plants, in line with the yellowing of leaves. Approximately 75% of chlorophyll was lost in saul1 leaves at a PFD < 40 μmol m−2 sec−1 (Figure 3a). Early senescence symptoms of the mutant also correlated with a decrease in photosynthetic activities. The photochemical efficiency of photosystem II, indicated by the optimum quantum yield Fv/Fm, declined with decreasing PFD (Figure 3b).

Leaf senescence is accompanied by changes in gene expression (Buchanan-Wollaston et al., 2003, 2005; Gepstein et al., 2003; Guo et al., 2004; Lin and Wu, 2004; Zentgraf et al., 2004). We analyzed the expression of chlorophyll biosynthetic and senescence-associated genes in saul1 mutants compared to wild-type under light conditions that resulted or not in the appearance of leaf yellowing in saul1 mutant plants (Figure 3c). We first looked at genes that encode components of chlorophyll biosynthesis. Expression of PORA, PORB and PORC encoding NADPH:protochlorophyllide oxidoreductases was reduced when saul1 mutants exhibited leaf yellowing at low PFD. This is in line with reduced chlorophyll biosynthesis, because an Arabidopsis porB porC double mutant was chlorophyll-deficient, leading to developmental arrest (Frick et al., 2003). In contrast, ACD1, encoding pheophorbide a oxygenase, was induced, supporting the observed chlorophyll degradation (Pruzinska et al., 2003; Yang et al., 2004). A slight induction was determined for At4g22920, which has recently been identified as Mendel’s I locus (Armstead et al., 2007). RNAi silencing of this gene caused a ‘stay-green’ phenotype in Arabidopsis plants (Armstead et al., 2007). Next we studied expression of the senescence marker gene SAG13. SAG13 mRNA was highly abundant only in saul1 mutant plants showing leaf senescence. Expression of the defense marker gene PR1, which is induced during leaf senescence (Robatzek and Somssich, 2001), was also increased in saul1 plants at low PFD. In addition, senescence is generally accompanied by autophagy for vacuolar bulk degradation of macromolecules. We assessed the regulation of genes encoding components of the autophagy machinery. All three genes tested, APG7, APG9 and ATG18A, were induced in saul1-1 plants with symptoms of senescence, suggesting that autophagy is activated in these mutants (Doelling et al., 2002; Hanaoka et al., 2002; Xiong et al., 2005). Furthermore, genes encoding signaling components that support senescence, such as SIRK (senescence-induced receptor kinase), WRKY6 and NAP, were induced in saul1 mutants at low PFD. Taken together, the described gene expression changes are in agreement with the visible senescence of saul1 mutants at low PFD.

Regulation of SAUL1 expression by development, day length and PFD

To test for SAUL1 promoter activity in leaves of various ages, we studied the expression pattern in SAUL1 promoter:GUS plants. We detected GUS staining in leaves at all developmental stages under short- and long-day conditions, indicating promoter activity during the entire lifecycle (Figure S2a). We did not observe significant differences in transcript abundance during the day under short-day conditions (data not shown) or long-day conditions (Figure S2b). RT-PCR experiments revealed no PFD-dependent changes in SAUL1 expression (Figure S2c). In wild-type plants that were subjected to darkness to induce senescence, the SAUL1 transcript was still present (Figure S2d). These data were in agreement with detection of SAUL1 transcripts in leaves of various developmental stages (Mudgil et al., 2004; Zimmermann et al., 2004). Our results on SAUL1 expression suggest that SAUL1 is not significantly regulated at the transcriptional level. We did also not observe regulation of SAUL1 transcript levels in Col wild-type leaves that were treated with ABA (Figure S2h). In addition to SAUL1 promoter activity in leaves, the SAUL1 promoter was active in the root vasculature and in guard cells (Figure S2e–g).

The mutation in SAUL1 affects ABA levels but not ABA signaling

During senescence, ABA levels are increased (Gepstein and Thimann, 1980). To assess the role of ABA signaling in the onset and progression of senescence, we sprayed wild-type plants and ABA signaling mutants with ABA. Whereas wild-type plants started to develop visible symptoms of leaf senescence after 1–2 weeks of ABA treatment, the ABA-insensitive signaling mutants abi1-1 and abi2-1 were not affected (Figure 4a). In contrast, the ABA-hypersensitive mutant era1-3 already showed early leaf senescence after 3 days of ABA treatment (Figure 4b). To find out whether ABA signaling is important for the saul1 senescence phenotype, we treated saul1 mutant and wild-type plants with ABA. All plants simultaneously initiated senescence, indicating that ABA signaling is not impaired by the mutation in SAUL1 (Figure 4c). Interestingly, ABA levels were changed in saul1 mutants. At low but not at high PFD, the leaves of saul1-1 mutants contained tenfold more ABA when compared to wild-type rosette leaves, suggesting that ABA biosynthesis was activated (Figure 4d). Increased ABA levels were also observed in saul1-2 mutants (Figure S3). Usually, the ABA content is closely linked to ethylene production, which promotes senescence. We therefore treated saul1-1 mutants with 5–20 μm aminoethoxyvinylglycine (AVG) to block ethylene biosynthesis. However, this treatment did not affect the senescence phenotype, indicating that ethylene production was not involved (Figure S1f). To support involvement of ABA in saul1 senescence, we crossed saul1-1 with abi1-1 mutants. The isolated saul1-1abi1-1 double mutants showed partial suppression of the saul1-1 phenotype (Figure 4e).

Figure 4.

 ABA signaling and biosynthesis in leaf senescence.
(a) Arabidopsis wild-type plants of ecotypes Columbia and Landsberg erecta, and abi1-1 and abi2-1 mutant plants were treated with ABA or mock solution for 1–2 weeks and scored for visible symptoms of leaf senescence.
(b) Columbia wild-type and era1-3 mutant plants were treated with ABA or mock solution for 3 days and scored for visible symptoms of leaf senescence.
(c) Senescence in saul1 mutant and wild-type plants that were grown on agar at a PFD of 90 μmol m−2 sec−1 and sprayed with ABA (bottom) or mock solution (top) for 5 days.
(d) ABA levels were measured immunologically in saul1 and wild-type plants that were subjected to low and high PFDs as indicated. Mean values ± SD from three independent experiments are shown. Student’s t-test analysis showed that the difference between ABA levels in saul1 mutants and wild-type at a PFD of 120 μmol m−2 sec−1 was not statistically significant, as indicated by the P value.
(e) Partial suppression of saul1-1 phenotypes in saul1-1 abi1-1 double mutants. Note that the double mutants were less developed compared to wild-type plants. Plants were grown for 12 days at a PFD of 20 μmol m−2 sec−1.

The AOδ protein is activated and stabilized in saul1 mutants

Activation of ABA biosynthesis is generally achieved by induction of the ABA biosynthetic genes NCED3 (9-cis-epoxycarotenoid dioxygenase 3) or AAO3 (Arabidopsis aldehyde oxidase 3) (Iuchi et al., 2001; Melhorn et al., 2008). In saul1 mutants, AAO3 expression was induced 2.3-fold at low PFD (P = 0.002), whereas the expression of NCED3 was not significantly changed (P = 0.08) (Figure 5a). We therefore focused our studies on AAO3, encoding the aldehyde oxidase isoform AOδ that catalyzes the last step in ABA biosynthesis, namely oxidation of abscisic aldehyde to ABA in rosette leaves (Seo et al., 2000a). Using native PAGE and in-gel activity, we were able to demonstrate that AOδ activity was increased in saul1 mutants at low but not high PFD (Figure 5b). AAO3 protein levels were also higher in young saul1 mutant seedlings at low PFD (Figure 5c). In adult saul1 mutant plants, activity and protein levels were increased when grown at a PFD of 250 μmol m−2 sec−1, but were not changed when grown at a PFD > 400 μmol m−2 sec−1 (Figure 5d). In saul1-1 35S-SAUL1 plant lines, which exhibited induced expression of SAUL1 but normal progression of senescence, the activity and protein levels returned to wild-type levels (Figure 5e). Apparently, the absence of SAUL1 leads to stabilization and thus accumulation of AAO3 at low PFD. Accumulation of AAO3 mRNA, AOδ activity and increased AAO3 protein levels were also observed during developmentally induced leaf senescence in wild-type plants, suggesting a general role for AAO3 in senescence (Figure 5f,g). Interestingly, when grown at low PFD, saul1-1 plants were indistinguishable from wild-type plants until the emergence of leaves 3 and 4. At that developmental stage, AAO3 was still absent from wild-type and saul1-1 plants (Figure 5h). However, AAO3 protein was detected in wild-type plants shortly after. While this background AAO3 level is not sufficient to trigger leaf senescence in wild-type plants, increased AAO3 protein levels coincided with the first symptoms of leaf senescence in saul1-1 plants (Figure 5h). The difference in AAO3 protein levels between wild-type and saul1-1 mutants became even more pronounced later in development (Figure 5c).

SAUL1 has E3 ubiquitin ligase activity and interacts with AAO3

To determine whether SAUL1 functions as E3 ubiquitin ligase, in vitro ubiquitination assays were performed. Human His-tagged E1 and E2 enzyme and biotinylated ubiquitin were used in these assays. Bacterial proteins present in the SAUL1–MBP and maltose binding protein (MBP) control extracts served as substrates for ubiquitination, respectively (Hatakeyama et al., 1997; Lorick et al., 1999). Detection of multiple high-molecular-mass proteins in the SAUL1–MBP extract in comparison to only few background signals in the control extract indicated the E3 ubiquitin ligase activity of SAUL1 (Figure 6a). In the absence of ubiquitin, E1 enzyme or E2 enzyme, such signals could not be detected (Figure 6b). The results showed that the U-box protein SAUL1 functions as a ubiquitin ligase. We next tested whether interactions occur between SAUL1 and AAO3. We labeled AAO3 with the trifunctional cross-linker Mts-Atf-biotin (Layer et al., 2007) and mixed it with SAUL1. Immunoblotting using a horseradish peroxidase–NeutrAvidinTM conjugate detected intense biotin labeling of SAUL1 (Figure 6c).This label transfer indicated that at least a transient interaction has occurred between AAO3 and SAUL1.

Figure 6.

 Ubiquitin ligase activity of SAUL1 and SAUL1–AAO3 interactions.
(a) In vitro ubiquitination assay using MBP–SAUL1 fusion proteins. Protein extracts were subjected to SDS–PAGE, and biotinylated ubiquitin was used for immunodetection of ubiquitin-labeled high-molecular-mass proteins. MBP was used in control experiments.
(b) In vitro ubiquitination assays were performed as in (a). Ubiquitin, the E1 enzyme or the E2 enzyme were omitted from the reactions as a negative control.
(c) Label transfer from AAO3 to SAUL1. Equal amounts of labeled AAO3 were used so that same amounts of Mts-Atf-biotin label were present. Immunoblotting detected the biotin label in AAO3 and truncated AAO3 proteins (left panel). When labeled AAO3 was incubated with SAUL1, an additional band that was absent from AAO3 and SAUL1 control samples was detected (arrow). The size of this band matched the size of SAUL1 as indicated by Coomassie staining (right panel).

The saul1 mutation induces expression of senescence-associated signaling components

Signaling components have been implicated in the induction of leaf senescence (Lim et al., 2007). In general, over-expression of these components causes early leaf senescence. Using real-time RT-PCR, we studied expression of the corresponding genes in saul1 mutant plants in comparison to their expression in wild-type plants. When plants were grown at high PFD, no differences in gene expression were observed (Figure 7a). However, the expression of SIRK, NAP, WRKY6, WRKY53, EDR1 (enhanced disease resistance 1) and AHK3 (Arabidopsis histidine kinase 3) was increased in saul1 mutants grown at low PFD, but EIN2 (ethylene insensitive 2) expression was not affected (Figure 7a). The highest induction was detected for SIRK (approximately 1000-fold), WRKY6 (approximately 9.5-fold) and NAP (approximately 4.5-fold). In view of the increased ABA levels detected in saul1 mutant plants under low PFD, ABA regulation of gene expression was determined in detached rosette leaves that were treated with ABA or mock solution. Expression of SIRK, WRKY6, EDR1 and AHK3 was not affected, but NAP and WRKY53 transcript levels increased upon ABA treatment, suggesting a role for NAP and WRKY53 in early signaling steps that lead to premature leaf senescence in saul1 mutant plants (Figure 7b).

Figure 7.

 Transcript levels of signaling components for leaf senescence in saul1 mutants and after ABA treatment.
(a) Real-time RT-PCR analyses of gene expression in wild-type and saul1 plants at the indicated PFD. The same plant material as in Figure 3(c) was analyzed. Expression in saul1 mutants at low PFD was set to 1 for easier comparison. Data represent mean values ± SD (n = 3).
(b) Quantitative analyses of gene expression after ABA treatment. Detached Col-0 rosette leaves were incubated in 50 μm ABA (black bars) or mock solution (white bars) and subjected to real-time RT-PCR. Data represent mean values ± SD from three independent ABA treatments.

Discussion

Protein degradation via the ubiquitin–26S proteasome pathway has key regulatory functions in plant development. It has been implicated in almost every developmental process from embryogenesis to generation of floral organs due to its central role in hormone and light signaling (Vierstra, 2003; Moon et al., 2004; Hoecker, 2005; Dreher and Callis, 2007). In this study, we have characterized the PUB protein SAUL1 as E3 ubiquitin ligase, and have shown that it prevents premature senescence under low light conditions. The leaves of saul1 mutants grown at low PFD exhibited visible yellowing accompanied by a decrease in chlorophyll content and the photochemical efficiency of photosystem II, and enhanced expression of senescence markers (Figures 1–3). The SAUL1 protein functions as an E3 ubiquitin ligase (Figure 6), and our data emphasize the role of ubiquitin-dependent protein degradation in the onset and progression of leaf senescence. In contrast to SAUL1, the F-box protein ORE9 appears to be a positive regulator of leaf senescence, because ore9 mutants showed a delay in leaf senescence. However, the ubiquitin ligase function of the ORE9-containing E3 complex remains to be shown (Woo et al., 2001).

PUB proteins comprise a family of 61 members, with important functions predominantly in plant defense (Azevedo et al., 2001; Dreher and Callis, 2007). When challenged with pathogens, plants increase the transcription of PUB-ARM E3 ligases. Among these, rice SPL11 is a negative regulator of plant cell death and defense, whereas tobacco CMPG1, tobacco ACRE276 and Arabidopsis PUB17 are required for defense (Zeng et al., 2004; Gonzalez-Lamothe et al., 2006; Yang et al., 2006). Other PUB proteins have been shown to participate in various hormone signaling pathways. Antisense inhibition of the predicted E3 ligase gene PHOR1 led to semi-dwarfism of potato plants that was accompanied by elevated gibberellin levels. Over-expression of PHOR1 produced plants with an enhanced response to gibberellin (Amador et al., 2001). There is evidence that the U-box E3 ligase CHIP is involved in ABA signaling. Over-expression of CHIP rendered Arabidopsis plants hypersensitive to ABA with respect to seed germination (Luo et al., 2006). Recently, a role for Arabidopsis PUB22 and PUB23 in the drought stress response has been shown (Cho et al., 2008). Here, we identified a PUB-ARM protein, SAUL1, that functions in the regulation of leaf senescence. It suppresses premature senescence of young seedlings at low PFD (Figures 1–3). Many known components of the signaling pathways that control senescence are positive regulators, as their knock-out results in delayed senescence but their over-expression leads to early senescence (Gan and Amasino, 1997; Thomas et al., 2003; Yoshida, 2003; Lim et al., 2007). In contrast, saul1 mutants exhibited early senescence in young seedlings with four leaves, and SAUL1 can thus be regarded as a negative regulator of senescence.

ABA increase may be a primary cause of leaf senescence in plants (Gepstein and Thimann, 1980). Indeed, we showed that ABA signaling is essential for onset of senescence after ABA treatment of Arabidopsis plants. Onset of senescence was accelerated in the ABA-hypersensitive signaling mutant era1-3, but suppressed in the ABA-insensitive mutants abi1-1 and abi2-1. In saul1 mutants, ABA signaling was normal, but ABA biosynthesis was increased at low PFD. The results of analysis of saul1-1 abi1-1 double mutants showing partial suppression of saul1-1 phenotypes strongly support the involvement of ABA in senescence (Figure 4). The last step of ABA biosynthesis, the conversion of abscisic aldehyde to ABA, is catalyzed by the aldehyde oxidase isoform AOδ encoded by AAO3. Expression of AAO3 mRNA is high in rosette leaves and increases upon dehydration (Seo et al., 2000b). Drought stress not only increases AAO3 transcript levels but also the activity of AOδ (Seo et al., 2000b; Bittner et al., 2001). Mutant analyses showed that AOδ plays important roles in ABA biosynthesis in rosette leaves and seeds, because aao3 mutants exhibited a wilty phenotype in rosettes and increased osmotolerance in seed germination (Seo et al., 2000a; Gonzalez-Guzman et al., 2004). High AAO3 expression was sufficient to induce stomatal closure (Melhorn et al., 2008). In ABA-accumulating saul1 mutant plants, AAO3 mRNA levels, AAO3 protein levels and AOδ activity were also increased, indicating that this enzyme has a key function in up-regulation of ABA synthesis in these plants (Figure 5a–d). Such an increase in AAO3 action was only observed at lower PFDs, irrespective of the developmental stage of the plants (compare Figures 2 and 5), suggesting that there is no need for SAUL1 action to suppress premature senescence at higher PFD. However, below a developmentally determined threshold of PFD, SAUL1 appears to be activated. We did not observe significant regulation of SAUL1 at the transcriptional level (Figure S2). In line with this, SAUL1 over-expression from the CaMV 35S promoter in saul1-1 plants did not result in delayed senescence. We conclude that SAUL1 function is regulated at the level of protein abundance or activity rather than transcription. Future research is required to unravel the molecular mechanisms that lead to activation of SAUL1 at low PFD and thus suppression of premature senescence. AAO3 protein accumulation was also observed during developmentally induced leaf senescence, indicating a general role for AAO3 in onset and progression of senescence (Figure 5f,g). Stabilization of AAO3 protein in the absence of SAUL1 in saul1 mutants at low PFD suggests that SAUL1 acts as E3 ubiquitin ligase either targeting AAO3 or components that regulate AAO3 transcription, stability and/or activity for proteolysis. As a prerequisite for interplay between SAUL1 and AAO3, the SAUL1 promoter is active in cells and tissues that harbor the AAO3 protein, such as root tips, root vasculature, leaf veins and guard cells (Figure S2a,e–g) (Koiwai et al., 2004). SAUL1-dependent AAO3 degradation may be required to prevent ABA accumulation and thus premature senescence in rosette leaves under low light conditions. In label transfer experiments, we showed that transient interactions between SAUL1 and AAO3 did indeed occur, thus supporting our hypothesis (Figure 6c). Strikingly, before showing symptoms of senescence, saul1 mutant plants develop normally under low light conditions until they reach the seedling stage with four leaves. This coincides with the absence of AOδ activity in 6-day-old seedlings (Seo et al., 2000b). Eight-day-old saul1-1 seedlings have a wild-type appearance, and the AAO3 protein is absent from wild-type and mutant plants. Shortly after, however, AAO3 can be detected in the wild-type and at an even higher level in saul1-1 plants. It is at this developmental stage that saul1-1 mutant plants start to show the first symptoms of leaf senescence (Figure 5h). Similar to saul1 mutants, up-regulation of ABA levels was also observed in Arabidopsis plants over-expressing the RING-H2 gene XERICO, which were more tolerant to drought (Ko et al., 2006). Although ubiquitin ligase activity has not yet been demonstrated for XERICO, these data also indicate regulation of ABA biosynthesis by the degradation of key biosynthetic enzymes or regulators of these enzymes.

Using transgenic Arabidopsis plants, it has been shown that strong expression of signaling components induces early senescence. SIRK, encoding senescence-induced receptor kinase, is specifically induced in senescing leaves. Transcriptional activation of SIRK depends on the WRKY6 transcription factor, which triggers necrosis in leaves, early flowering and reduced apical dominance when over-expressed (Robatzek and Somssich, 2002). Over-expression of the EDR1 kinase domain (enhanced disease resistance 1) enhanced ethylene-induced senescence (Tang et al., 2005). Similarly, over-expression of NAP, an Arabidopsis gene encoding a NAC family transcription factor, or of WRKY53 caused premature senescence (Miao et al., 2004; Guo and Gan, 2006). All these genes (SIRK, WRKY6, EDR1, WRKY53 and NAP) were induced in saul1 mutant plants showing yellowing of leaves under low light (Figure 7a). In contrast, ABA treatment of detached leaves resulted only in the induction of NAP and WRKY53 (Figure 7b), suggesting a role for these transcriptional regulators high in the signaling cascade that leads to early senescence in saul1 mutants at low PFD.

Our results support a model for regulation of the onset and progression of leaf senescence by the level of ABA (which depends on AAO3 activity) and by the E3 ubiquitin ligase SAUL1, which, when present and active, participates in targeting AAO3 for degradation via the ubiquitin–26S proteasome pathway. Transient interactions do indeed occur between SAUL1 and AAO3. In saul1 mutants, AAO3 and ABA accumulation resulted in premature senescence under low light conditions, indicating that SAUL1 is a negative regulator of leaf senescence in wild-type plants. Detailed kinetic analyses of SAUL1 and AAO3 action, ABA biosynthesis and expression of regulatory factors such as NAP and WRKY53 will help to determine the timing of signaling events that suppress premature senescence.

Experimental procedures

Plant material, strains, growth conditions and treatments

Arabidopsis thaliana wild-type and transgenic plants were grown on potting soil at 21°C under short-day (8 h light/16 h dark) or long-day conditions (16 h light/8 h dark). For growth on agar, seeds were surface-sterilized and sown on Petri dishes containing Murashige and Skoog salts pH 5.7, 0.8% phytoagar (Duchefa Biochemie, http://www.duchefa.com) and 1% sucrose. The photon flux density (PFD) was measured using an LI-250A light meter combined with a Quantum sensor (LI-COR Biosciences, http://www.licor.com). For ABA treatment, detached rosette leaves were submerged for 4 h in 50 μm ABA aqueous solution or in mock solution containing an equal amount of methanol (Sigma-Aldrich, http://www.sigmaaldrich.com/) and 0.01% Tween-20. Identical solutions were also used for spraying plants with ABA. For cloning in Escherichia coli, we used strain DH5α (Hanahan, 1983). Transformation of Arabidopsis was performed using Agrobacterium tumefaciens strain GV3101 (Holsters et al., 1980).

Cloning of DNA constructs and transformation of Arabidopsis

For generation of the SAUL1 promoter:reporter gene construct, we isolated the SAUL1 promoter from A. thaliana ecotype Columbia. We amplified a 1458 bp SphI/PciI fragment using primers 5′-CTGGCATGCTCCTTTTTAATCCCAGATTC-3′ and 5′-GTCACATGTCTTAATCCTTCAAACTCACTG-3′. The fragment was cloned into pCR®-Blunt II-TOPO® (Invitrogen, http://www.invitrogen.com/) and verified by DNA sequencing. The insert was cloned in front of the GUS reporter gene in the vector pAF6, a pUC19-based plasmid harboring the GUS gene. The promoter:GUS fragment (containing 1322 bases of the SAUL1 promoter) was excised and cloned into pAF16 (Stadler et al., 2005) via HindIII and SacI. For the complementation construct CaMV 35S promoter:SAUL1, we amplified an XhoI/PacI fragment of the SAUL1 cDNA using primers 5′-CTACTCGAGATGGTTGGAAGCTCGGAT-3′ and 5′-GTTTTAATTAACTATGCGATGTTTGGGAATATAC-3′. The fragment was cloned into pCR®-Blunt II-TOPO® (Invitrogen), sequenced, and cloned into pBA002 (Kost et al., 1998) via XhoI and PacI. The generated plasmids were used for transformation of Arabidopsis wild-type and saul1-1 plants by floral dip (Clough and Bent, 1998). Transformants were obtained by Basta selection (Bayer Crop Science, http://www.bayercropscience.de). For generation of the MBP–SAUL1 fusion construct, the SAUL1 coding sequence was amplified using primers 5′- CTACTCGAGATGGTTGGAAGCTCGGATG-3′ and 5′- GATACGCGTCTATGCGATGTTTGGGAATATAC-3′ and cloned into pCR®-Blunt II-TOPO® (Invitrogen). An EcoRI fragment was excised and cloned into pMAL-c2X (New England Biolabs, http://www.neb.com).

Staining for GUS activity, RNA isolation, RT-PCR and real-time RT-PCR analysis

These experiments were performed as described previously (Raab et al., 2006). Primers used for RT-PCR analyses are listed in Table S1. Relative expression levels are given in Figures 1e, 5a, 5f and 7.

Isolation and characterization of homozygous T-DNA insertion lines

Homozygous saul1 mutants were isolated from seeds of segregating T3 lines of T-DNA insertion lines SALK_063974 and SALK_076799 (Alonso et al., 2003). After selfing the original plants, the positions of the T-DNA insertions were determined by sequencing fragments amplified using T-DNA left border primer LBa (5′-TGGTTCACGTAGTGGGCCATCG-3′) and SAUL1-specific primers (5′-GCAAATGAAGTAGAGAGCG-3′ and 5′-GAAGCATTCATCTGCCC-3′).

Determination of chlorophyll content and pulse amplitude modulated fluorometry

To determine chlorophyll contents, leaf tissue was ground, and 80% acetone was added to a final volume of 1.5 mL. Samples were centrifuged at 3000 g for 2 min before measuring absorbance photometrically. Fluorescence in leaves was measured using an IMAGING-PAM fluorometer (Walz, http://www.walz.com). The Fv/Fm ratio was determined in dark-adapted plants using standard protocols.

ABA analysis, determination of aldehyde oxidase activity and AAO3 protein level

In freeze-dried plant material, ABA levels were determined as described previously (Becker et al., 2003). For detection of aldehyde oxidase activity, plant samples were squeezed at 4°C in two volumes of a solution containing 100 mm potassium phosphate, 2.5 mm EDTA and 5 mm dithiothreitol, pH 7.5, and sonicated. After centrifugation at 4°C and 20 000 g for 20 min, 50 μg of crude protein extracts were subjected to native PAGE on a 7.5% gel, and aldehyde oxidase activity was detected using indole-3-carboxaldehyde and 1-naphthaldehyde as substrates (Koshiba et al., 1996). For Western blot analysis of AAO3 protein abundance, crude extracts containing 50 μg of proteins were subjected to native PAGE on a 7.5% gel. Detection of AAO3 (and AAO2) protein was performed using polyclonal rabbit anti-AAO3 antibodies and anti-rabbit alkaline phosphatase conjugated secondary antibody. Protein bands were visualized using 5-bromo-4-chloro-3-indolyl phosphate and nitroblue tetrazolium as substrates according to standard protocols.

In vitro ubiquitination assay

The MBP–SAUL1 construct and the pMAL-c2 vector were transformed into E. coli strain DH5α. To express proteins, overnight cultures were diluted 1:100 in a total volume of 250 mL. Induction of proteins was performed using isopropyl-β-D-thiogalacto pyranoside for 2 h at 37°C. Following induction, cells were harvested by centrifugation at 4°C and 4000 g for 2 min, resuspended in 3 mL 1× PBS (140 mm NaCl, 2.7 mm KCl, 10 mm Na2HPO4, 1.8 mm KH2PO4) to which protease inhibitor (Roche Diagnostics, http://www.roche.com) had been added, and sonicated. After centrifugation (4°C and 9000 g for 30 min) of the obtained cell lysates, supernatants were used for in vitro ubiquitination assays. Reactions were performed in a total volume of 50 μL using components of the BIOMOL ubiquitin conjugation kit (BIOMOL, http://www.biomol.com). Each reaction contained 0.1 μm human E1, 2.5 μm E2 (UbcH5b, His6-tagged), 2.5 μm biotinylated ubiquitin, 1× ubiquitination buffer, 1 mm dithiothreitol, 1× ATP regeneration solution and 50 μg lysate as indicated. The reactions were incubated for 2 h at 37°C, and the reaction was stopped by adding 50 μL 2× non-reducing loading buffer. Protein samples were resolved on 10% SDS–PAGE, followed by gel-blot analysis using ExtrAvidin® peroxidase (Sigma-Aldrich).

Biotin-label transfer between AAO3 and SAUL1

Expression and purification of recombinant AAO3 was performed essentially as described for recombinant AtXDH1 (Hesberg et al., 2004; Koiwai et al., 2004). For heterologous expression of SAUL1 in E. coli DL41, full-length SAUL1 cDNA was cloned into pQE80 (Qiagen, http://www.qiagen.com/) between the BamHI and SalI restriction sites. Cells carrying the pQE80/SAUL1 construct were grown aerobically in LB medium containing 100 μg mL−1 ampicillin at 37°C and 150 rpm to an absorbance at 600 nm of 0.3–0.4 before induction with 0.5 mm isopropyl-β-d-thiogalactopyranoside. After induction, cells were grown for further 6 h at 30°C. Cells were centrifuged at 3000 g for 5 min at 4°C and 25 000 g for 20 min, resuspended in lysis buffer (50 mm sodium phosphate, pH 8.0, containing 200 mm sodium chloride, 10 mm imidazole and 5% glycerol), and broken by several passages through a French pressure cell followed by sonication for 5 min. After centrifugation, His6-tagged SAUL1 protein was purified using a nickel-nitrilotriacetic acid superflow matrix (Qiagen) under native conditions at 4°C according to the manufacturer’s instructions, and eluted in elution buffer (50 mm sodium phosphate, pH 8.0, containing 300 mm sodium chloride, 150 mm imidazole and 10% glycerol). After 7.5% SDS–PAGE and Coomassie staining, the fractions with highest purity of SAUL1 and AAO3 were chosen for interaction experiments, which were essentially performed as described by Layer et al. (2007). AAO3 was labelled with the Mts-Atf-biotin label (2-[N2-(Y-azido-2,3,5,6-tetrafluoro-beuzoyl)-N6-(6-biotinamidocaproyl)-L-lysinyl] ethyl methane-thiosulfonate (Pierce/Perbio, http://www.piercenet.com) prior to co-incubation of equimolar amounts of SAUL1 (11.1 μg) and AAO3 (21.6 μg). After 7.5% SDS–PAGE and gel blotting, transfer of the biotin label from AAO3 to SAUL1 was detected using NeutrAvidin™-conjugated horseradish peroxidase (1:35 000 dilution; Pierce/Perbio) and an enhanced chemoluminiscence system (Amersham/GE Healthcare, http://www5.amershambiosciences.com/).

Acknowledgements

We thank Norbert Sauer, Ulrich Hammes, Ruth Stadler, Petra Dietrich, Uwe Sonnewald, Mitsunori Seo and Barbara Niemeyer for discussions, and Stefanie Geigner and Rebecca Günther for technical assistance. This work was funded by Deutsche Forschungsgemeinschaft grants SFB473TPC11 and HO2234/4-1 to S.H.

Ancillary