An ubiquitin-like protein SDE2 negatively affects sucrose-induced anthocyanin biosynthesis in Arabidopsis.

Anthocyanin biosynthesis is regulated by a conserved transcriptional MBW complex composed of MYB, bHLH and WD40 subunits. However, molecular mechanisms underlying transcriptional regulation of these MBW subunits remain largely elusive. In this study, we isolated an Arabidopsis mutant that displays a constitutive red color in aboveground tissues with retarded growth phenotypes. In the presence of sucrose, the mutant accumulates more than 3-fold anthocyanins of the wild type (WT), but cannot produce anthocyanins as WT in the absence of sucrose. Map-based cloning results demonstrated that the mutation occurs in the locus At4G01000, which encodes a conserved nuclear-localized ubiquitin-like (UBL) superfamily protein, silencing defective 2 (SDE2), in eukaryotes. SDE2 is ubiquitously expressed in various tissues. In the sucrose-induced anthocyanin biosynthesis, SDE2 expression was not responded to sucrose treatment at the early stage but was enhanced at the late stage. SDE2 mutations result in up-regulation of anthocyanin biosynthetic and regulatory genes. Yeast-two hybrid analysis indicated that SDE2 has no direct interaction with the MYB transcription factor PAP1 and bHLH factor TT8, indicating that SDE2 is a indirect factor to affect anthocyanin accumulation. Taking together, our data suggest that SDE2 may play a role in finely coordinating anthocyanin biosynthesis with other biological processes.


Introduction
Anthocyanins are widely distributed secondary metabolites throughout the plant kingdom [1]. They protect plants from UV-B irradiation, endow plants with red to blue colors to attract pollinators and seed dispersers, and function as antimicrobial agents against pathogen infection [2,3]. The visible anthocyanin pigments in plants also provide an excellent platform for the study of molecular mechanisms underlying transcriptional regulation of genes [4,5]. Therefore, anthocyanin biosynthesis is intensively studied in many plant species, including the model species Arabidopsis thaliana [6,7].
Anthocyanin biosynthesis is regulated by a complicated interplay between internal and external signals such as carbohydrates, phytohormones, light, and temperature [19]. Sucrose has been widely reported to induce anthocyanin biosynthesis [16,20]. Phytohormones such as Jasmonic acid (JA), Gibberellic acid (GA) and cytokinin are also important internal signals to affect anthocyanin biosynthesis [21,22]. None of the phytohormones could significantly affect the expression of anthocyanin biosynthetic genes unless sucrose was applied concomitantly [21,23]. Namely, a certain level of sucrose is a prerequisite for the action of phytohormones with regard to regulation of anthocyanin biosynthesis. The expression of PAP1 and PAP2 were dramatically up-regulated in seedlings grown on the media containing sucrose [20,21]. Recently, ANGUSTIFOLIA3-YODA gene cascade was reported to induce anthocyanin accumulation by regulating sucrose levels via the change of invertase activity [24]. In addition, the transcription factor ANAC032 was demonstrated to repress anthocyanin biosynthesis under the condition of high sucrose levels [25]. In general, the molecular mechanism underlying sucrose-induced anthocyanin biosynthesis remains largely unclear.
In this study, by using sucrose to induce anthocyanin accumulation in Arabidopsis seedlings, we identified a mutant with hyperaccumulation of anthocyanins. The mutated gene encodes a conserved ubiquitin-like (UBL) superfamily protein in eukaryotes. Map-based cloning and bioinformatic analysis showed that the identified gene is the ortholog of the yeast SILENCING DEFECTIVE 2 (SDE2), which is essential for telomeric silencing and genomic stability [26]. Thus, we named the mutant as sde2. The ubiquitin and ubiquitin-like modifiers are well characterized to covalently attach to target proteins via a lysine side chain in the 26S proteasome system [27]. In contrast, the proteins bearing a ubiquitinlike (UBL) domain are involved in regulating ubiquitin signaling. For example, the UBL protein Rad23 acts as a shuttle factor to carry an ubiquitinated substrate to the 26S proteasome [28]; the human SDE2 is required for the monoubiquitination level of PCNA (proliferating cell nuclear antigen), which is induced by ultraviolet DNA damage and counteracts replication stress [29]. In Arabidopsis, SDE2 is highly expressed in anthocyanin-accumulated tissues. Loss-of-function mutations in SDE2 result in a significant increase in the expression of several key regulatory genes, such as PAP1/ PAP2, TT8 and TTG1, and all examined anthocyanin biosynthetic genes. Our findings provide a new insight into the transcriptional regulation of the MBW activity by SDE2, which may finely tune the coordination of anthocyanin biosynthesis with other biological processes in vivo.

Plant materials and growth conditions
All Arabidopsis thaliana plants used in this study were the Columbia-0 (Col-0) or Landsberg erecta (Ler) background. The transfer DNA (T-DNA) insertion pools from CS75001 to CS75300 were obtained from the Arabidopsis Biological Resource Center (ABRC) and screened for mutants with anthocyanin overaccumulation. The sde2-1 (CS75072-2) mutant was isolated from the pool CS75072, and the sde2-2 (Salk_114391C) mutant was identified from another T-DNA (pROK2) pool ordered from the ABRC. All these mutants were in the Col-0 background. The surfacesterilized seeds were treated at 4°C in the dark for 3 days and then germinated on the half strength Murashige and Skoog (MS) medium (pH 5.7, Sigma-Aldrich) containing 0.7% agar and 1% sucrose at 22°C with a light intensity of 100 lmol of photons m À2 s À1 in long-day conditions (16-h light/8-h dark). Seven-day-old seedlings were transferred to soil and grown in a greenhouse under the same conditions as seed germination.
For screening anthocyanin-accumulated mutants, 3-day-old T-DNA insertion seedlings and wild-type (Col-0) seedlings grown in half MS liquid media without sucrose on a shaker at 120 rpm were transferred to half MS solution with 3% (w/v) sucrose to induce anthocyanin biosynthesis for additional 2 days. Then, seedlings were transferred into sucrose-free half MS solution for another additional 2 days. The mutants with visible red or purple color were selected to grow in the greenhouse.

TAIL-PCR, map-based cloning and mutant identification
TAIL-PCR was performed as previously described [30], and the primers used in this study were listed in Table S1. Map-based cloning was performed according to the report by Jander et al. In brief, the sde2 mutant was crossed with Ler to generate an F2 mapping population. SSLP makers (Table S1, online) were designed based on lndel polymorphisms between Col and Ler ecotypes to map the gene. The information is available at TAIR (http://www.arabidopsis.org). The gene was roughly mapped by bulked-segregant analysis using a mixed pool of 50 F2 seedlings and was finely mapped using a population of about 200 F2 plants. The specific primers were used to identify sde2-1 and sde2-2 mutants.

Plasmid construction and plant transformation
Plasmids were constructed using the Gateway system (Invitrogen). For genetic complementation, the SDE2 genomic DNA including the 2-kb promoter region upstream of the ATG start codon, three exons and two introns was amplified and inserted into pENTR/SD/D-TOPO (Invitrogen, K2420-20), and then transferred into the pGWB1 binary vector (Research Institute of Molecular Genetics, Shimane University, Japan). For the construct of GUS expression, the 2 kb promoter fragment was amplified and cloned into pENTR/SD/D-TOPO (Invitrogen, K2420-20), and subsequently into the pGWB3 binary vector. The constructs were transformed into the Agrobacterium tumefaciens GV3101 strain and introduced into Arabidopsis plants by the floral dip method [31]. Transgenic plants were screened by hygromycin resistance.

Gene expression analysis
RNA extraction and cDNA preparation were performed as described above. The SDE2 expression was amplified by RT-PCR with the gene-specific primers (Table S1, online). Quantitative RT-PCR was used to analyze the relative expression levels of genes involved in anthocyanin biosynthesis, and ACTIN2 (At3g18780) was used as an internal control [23] . Normalized relative expression was calculated by the DDCt method. The amplifications were performed by Applied Biosystems StepOnePlus TM Real-Time PCR system with a total volume of 20 lL of SYBR Premix Ex Taq TM (TakaRa, RR402A). Two biological repeats were performed. The primers used were described in Table S1.

Anthocyanin quantification and profiling by HPLC/DAD (Diode Array Detector)
Frozen seedlings were homogenized in 1 mL of extraction solution (methanol: HCl: H 2 O = 79:1:20) per mg of fresh weigh overnight at 4°C. After centrifugation at 12,000 g, the 10 lL supernatants were applied to HPLC/DAD (Agilent 1260) for anthocyanin quantification and profiling. HPLC was performed on a C18 column (3.5 lm, 4.6 Â 100 mm Agilent) at a flow rate of 0.7 mL min À1 , elution gradient with solvent A (H 2 O-0.1% CF 3 COOH) and solvent B [CH 3 CN] and the following elution gradient (92% to 78% A from 0 to 30 min, and 78% to 92% A from 30 to 35 min, isocratic at 92% A from 35 to 40 min to equilibrate the column for the next injection). DAD was used for detection of UV-visible absorption of 520 nm. For anthocyanin quantification and identity, significant peaks separated by HPLC were collected by a fraction collector (Agilent Technology) and then applied to LC-ESI/Q-TOF/MS as described previously.

Y2H assay
Yeast AH109 cells were cotransformed with specific bait and prey constructs through the LiCl-PEG method according to the manufacturer's instructions (Clontech). The transformants were selected on the media lack of Leucine and threonine (-LT). The interactions were tested on the media without leucine, threonine and histidine (LTH) or without LTH plus 3-amino-1,2,4-triazole (3-AT). For screening SDE2-interacting proteins, the ORF of SDE2 was amplified and finally cloned into pDEST32 to generate the bait. The bait was co-transformed with a prey, a library of Arabidopsis encoding sequences cloned into pDEST22. Plasmids were extracted from the positive clones using the TIANprep Yeast plasmid DNA kit (TIANGEN, DP112), and sequenced to know the candidate genes encoding SDE2-interacting proteins.

Statistical analyses
Statistical analyses were performed using Student's t-test embedded in the Microsoft Excel software. Only the value of P < 0.05 was designated as statistically significant.

Screening for mutants with hyperaccumulation of anthocyanins
To isolate new genes involved in anthocyanin metabolism and regulation, we established a simple system to screen mutants with hyperaccumulation of anthocyanins using liquid-cultured Arabidopsis seedlings. Based on the sucrose-specific induction of anthocyanin accumulation in Arabidopsis seedlings [20], anthocyanin biosynthesis is first induced by sucrose for 2 days [16], and then sucrose is removed to initiate anthocyanin degradation.
Usually, anthocyanins in wild-type (WT) seedlings are completely degraded 2 days after the removal of sugar. The red phenotype of the isolated mutants can be attributed to either up-regulated biosynthesis or down-regulated degradation of anthocyanins. Here, we characterized a mutant that exhibits a constitutive red color in aboveground tissues, and here was designated as sde2-1 (silencing defective 2). The mutant was identified in the T-DNA insertion mutant pool, CS75072, from the ABRC. Anthocyanins apparently accumulated in cotyledons and hypocotyls of the mutant, compared with those of the WT (Fig. 1a). In the absence of sucrose, both WT and sde2 seedlings did not produce anthocyanins (Fig. S1a, online). Interestingly, when treated with 300 mmol/L mannitol, anthocyanin accumulated apparently in WT cotyledons but not in sde2 cotyledons (Fig. S1b, online). These data suggesting that sucrose-specific induction of anthocyanin accumulation is not altered and also not caused by osmotic stress in sde2. In addition, sde2 plants had red/purple inflorescence stems, whereas WT plants had green ones (Fig. 1b). The mutant is also small in size of the whole body and organs (Fig. S2, online). For example, the silique length and plant height of the mutant was about half and 70% of the WT, respectively. These results suggest that SDE2 functions not only in anthocyanin accumulation but also in plant growth and development.
To know whether anthocyanin hyperaccumulation of the mutant is caused by enhanced biosynthesis or reduced degradation, we analyzed anthocyanin content in a time-dependent course. Our results showed that the sde2-1 mutant accumulated more than 3-fold anthocyanins of WT, whereas there was no significant difference in the rate of anthocyanin degradation between the mutant and WT (Fig. 1c). In addition, we analyzed the components of anthocyanins using high-performance liquid chromatography (HPLC). As shown in Fig. 1d, the same pattern of anthocyanin components was detected in both WT and mutant seedlings grown in sucrose-contained half MS solution, although the value of the major peaks from a1 to a6 differed between the WT and mutant. These results indicated that anthocyanin hyperaccumulation in sde2-1 is due to enhanced biosynthesis but not due to reduced degradation.

SDE2 encodes an ubiqutin-like protein
Since sde2-1 was created by T-DNA insertion mutagenesis, we used Thermal Asymmetric Interlaced PCR (TAIL-PCR) to identify the mutated gene [30]. Unfortunately, TAIL-PCR analysis showed that there was no any T-DNA insertion in the mutant. We then used the map-based cloning method to isolate the gene. The F2 mapping population was generated by crossing sde2 with Landsberg erecta (Ler). Using simple sequence length polymorphism (SSLP) markers, we mapped the gene to the region between T18A10 and F15P23 makers on chromosome 4 (Fig. 2a). DNA sequencing revealed that there was a deletion of 17 nucleotides at the second exon of the gene At4g01000, which encodes an ubiqutin-like superfamily protein. The deletion leads to the formation of a premature stop codon. Phylogenetic analysis of the gene locus At4G01000 via TAIR website (http://www.arabidopsis.org/) demonstrated that SDE2 is conserved in eukaryotes. There are two domains, namely the N-terminal UBL domain and the C-terminal SDE2 domain in SDE2. In addition, the Arabidopsis genome contains another homolog (At3g06455) of SDE2, sharing about 50% identity at the protein level. However, their functions remain unclear.
To confirm the map-based cloning result, we transformed the mutant with the SDE2 cDNA driven by its native promoter. As expected, transgenic plants expressing SDE2 complemented all phenotypes including anthocyanin accumulation of the mutant, indicating that the map-based cloning result is correct (Fig. 2b,  c). In addition, we identified another T-DNA insertion line (Salk_114391C), named sde2-2, from seed stocks of the ABRC (Fig. 2d). sde2-2 displayed the same phenotype as sde2-1 (Fig. 2e). Taken together, our data suggest that loss-of-function mutations in SDE2 leads to over-accumulation of anthocyanins.

SDE2 is ubiquitously expressed in various tissues
To examine expression patterns of SDE2, we generated transgenic lines expressing the b-glucuronidase (GUS) gene driven by the native promoter. GUS activity analysis showed that the SDE2 promoter was active in almost all tissues in 6-day-old young seedlings and 7-week-old mature plants (Fig. 3a-h). The activity of the SDE2 promoter was strong in cotyledons and emerging leaves (Fig. 3a), root tips (Fig. 3b), inflorescences (Fig. 3 g), apical and basal parts of siliques (Fig. 3 h), while moderate in hypocotyls (Fig. 3a), leaves (Fig. 3d, e), and internodes (Fig. 3f). However, the GUS activity was not detected in trichomes (Fig. 3c). We also examined tissue-specific expression patterns of SDE2 using RT-PCR. Consistent with the results from GUS activity assays, SDE2 mRNA accumulated in all examined tissues, especially in inflorescences (Fig. 3i). Thus, the SDE2 expression patterns are in agreement with the multiple phenotypes of sde2 in the smaller size of the whole body and organs (Fig. S2 online, Fig. 2c, d), suggesting that SDE2 plays important roles in diverse physiological processes.
Since SDE2 is involved in sucrose-induced anthocyanin biosynthesis, we examined whether SDE2 expression is regulated by sucrose. Quantitative PCR analysis demonstrated that SDE2 expression was moderately inhibited by sugar within 2 h treatment, and maintained at a relatively stable level before 5 h treatment (Fig. 3j). This was in contrast to the expression of DFR and PAP1, whose levels were increased rapidly after sugar treatment (Fig. 3j). Sugar-induction of SDE2 expression was observed 10 h after sucrose treatment (Fig. 3j). Taken together, our data suggest that SDE2 is expressed in various tissues, and responds to sugar differentially in a time-dependent manner.

SDE2 mutations lead to enhanced expression of genes involved in anthocyanin biosynthesis
The phenotype of sde2 in over-accumulation of anthocyanins promoted us to examine the effect of SDE2 on expression of the genes involved in anthocyanin biosynthesis and regulation. We first investigated expression of a number of the regulatory genes using quantitative real-time PCR (qPCR). These genes include TTG1, PAP1/PAP2, TT8/GL3/EGL3, and MYBL2/CPC, which encode WD40 protein, R2R3-MYB transcription factors, bHLH transcription factors, and R3-MYB repressors, respectively. qPCR results showed that transcriptional levels of TTG1, PAP1, PAP2, and TT8 were significantly up-regulated while those of MYBL2 and GL3 were downregulated in sde2-1, compared with those in WT seedlings (Fig. 4a). There was no significant difference in the expression of CPC and EGL3 between sde2-1 and WT (Fig. 4a).
Taken together, our data indicate that over-accumulation of anthocyanins in sde2 is associated with higher expression levels of anthocyanin biosynthetic and regulatory genes, suggesting that SDE2 negatively affects anthocyanin biosynthesis.

SDE2 is localized in nucleus and indirectly regulates expression of PAP1/PAP2 and TT8
To study molecular mechanisms underlying SDE2-regulated expression of PAP1, PAP2, and TT8, we firstly examined the subcellular localization of SDE2 in Arabidopsis. Transiently expression of SDE2-EYFP fusion protein in Arabidopsis protoplasts showed that the GFP signal was exclusively localized in the nucleus (Fig. 5). Because SDE2 encodes an ubiqutin-like superfamily protein, we assumed two possibilities for SDE2 action: one is that SDE2 interacts with PAP1/PAP2 and/or TT8 to repress their transcriptional (c) Relative expression levels of TT19 and genes involved in anthocyanin modification in WT and sde2-1 seedlings by qRT-PCR. mRNA abundances were analyzed in 5-day-old seedlings grown in 3% sucrose-contained half MS media. ACTIN2 was used as an internal control; error bars represent ± SD (n = 3). The expression levels of genes in WT seedlings were set to 1. * P < 0.05; **, P < 0.01 (Student's t-test); Two biological replicates were analyzed with similar results. activation, the other is that SDE2 binds to other anthocyanin regulators and then inhibits the expression of PAP1/PAP2 and TT8. To test these two hypotheses, we first used the yeast two-hybrid (Y2H) assay to see whether PAP1/PAP2 and TT8 could interact with SDE2. As PAP1 and PAP2 are both R2R3-MYB transcription factors with redundant function in anthocyanin biosynthesis and share high sequence similarity [8], we examined interaction between PAP1 and SDE2. Our data showed that SDE2 did not interact with PAP1 or TT8 in yeast (Fig. S3). Then, we screened for SDE2interacting proteins using Y2H, and identified some potential binding proteins, in which nucleocytoplasmic transport (NTF2) and DNA-binding transcription factor (ASIL2) were confirmed to interact with SDE2 in yeast (Fig. S3). Taken together, our results suggest that SDE2 indirectly regulates expression of PAP1/PAP2 and TT8.

SDE2 negatively affects sucrose-induced anthocyanin biosynthesis
In this study, we identified a new protein SDE2 affecting anthocyanin biosynthesis (Fig. 2). It is interesting that overaccumulation of anthocyanins in loss-of-function sde2 mutants is sucrose specific and not attributed to osmotic stress. SDE2 mutations lead to increased expression of many anthocyanin biosynthetic genes in both early and later stages of the anthocyanin biosynthetic pathway (Fig. 4). In the process of sucrose-induced anthocyanin biosynthesis, SDE2 expression was not responded to sucrose at the early stage but was induced by sucrose at the late stage (Fig. 3j). The degree of the induction by sucrose is within 2-to 3-folds. Such an expression pattern may be helpful for plants to rapidly accumulate anthocyanins as well as to prevent from synthesizing too much anthocyanin, and to better adapt to environmental changes.

Possible mechanisms of SDE2-regulated anthocyanin biosynthesis
The SDE2 gene was first identified as a novel nuclear protein that is essential for telomeric silencing and genomic stability in the fission yeast [26]. Recently, SDE2 was shown to be processed at the first di-glycine motif by the ubiquitin-specific proteases Ubp5 and Ubp15, and the activated C-terminal part of SDE2 was subsequently incorporated into spliceosomes to ensure proper splicing of certain pre-mRNAs in fission yeast [39]. In human cells, SDE2 was also involved in protecting genomic integrity [29]. Likewise, SDE2 is cleaved at a diglycine motif of the UBL domain, and the cleaved SDE2 negatively regulates ultraviolet damageinducible monoubiquitination of PCNA, which plays a key role in coordinating DNA repair against replication-clocking lesions [29]. In this study, we found that SDE2 acted as a negative factor for expression of key anthocyanin regulatory genes such as PAP1, PAP2 and TT8. In yeast and humans, SDE2 has been demonstrated to be a regulatory component in the Ubiquitin/proteasomemediated proteolysis (UPP) pathway. Thus, we propose that SDE2 affects anthocyanin biosynthesis probably via the UPP pathway. Accumulating evidences have demonstrated that UPP dysfunction results in pleiotropic vegetative and reproductive phenotypes, including anthocyanin accumulation [17,40]. R2R3-MYB transcription factors PAP1 and PAP2 are degraded in the dark, where anthocyanin biosynthesis is repressed. The degradation of PAP1 and PAP2 is dependent on their interaction with COP1/SPA ubiquitin ligase. The COP1/SPA complex affects PAP1 and PAP2 both transcriptionally and posttranslationally [17]. ARABIDOPOSIS SKP1 HOMOLOGUE 1 (ASK1) gene encodes a subunit of a SCF (Skp, Cullin, F-box) ubiquitin ligase. ASK1 could physically interact with the above mentioned sucrose-repressed KIN10 [41], a negative regulator of PAP1 expression. Interestingly, SKP1-like protein ASK21, another component of the SCF ubiquitin ligase complex, whose mutation results in increased anthocyanin content [42], was identified as potential upstream transcriptional regulator of PAP1 [43]. In addition, the intrinsic 26S proteasome subunit REGULATORY PARTICLE NON-ATPASE 10 (RPN10), one major ubiquitin receptor, was involved in diverse physiological processes. Similar to sde2 mutants (Fig. 1b), rpn10-2 plants accumulate anthocyanins on the top floral stem [40]. However, we did not detect any interaction between SDE2 and PAP1/PAP2 or TT8 (Fig. S3). Future studies will be needed to unravel whether SDE2 can bind to COP1/SPA, ASK1/21, KIN10, and/or RPN10, and function in the same pathway to regulate the stability of PAP1/PAP2.

Pleiotropic phenotypes caused by SDE2 mutations
The SDE2 promoter activity was detected in almost all tissues. This is consistent with the results that SDE2 mRNA is ubiquitously expressed (Fig. 3). Subcellular localization analysis showed that SDE2 protein is localized in the nucleus (Fig. 5). Y2H screening of interacting proteins of SDE2 reveals that SDE2 could interact with various factors. These proteins are supposed to be involved in various physiological processes, such as protein ubiquitination and transport, metabolism, transcriptional regulation, and even chloroplast development. All these results are in agreement with the pleiotropic growth and developmental phenotypes caused by SDE2 mutations (Fig. S2). Thus, the roles of SDE2 in other processes will be an interesting topic to be investigated in the future.

Author contributions
Y.L., Y.X. Q. C. and J.H. designed the experiments, analyzed data, and wrote the article. Y.L. and Y.X. performed majority experiments, H.T. analyzed the anthocyanin contents.

Conflict of interest
The authors declare that they have no conflict of interest. the pictures of EYFP, chlorophyll and DAPI were merged; DIC (e), differential interference contrast. Bars = 10 lm.