The U-box/ARM E3 ligase PUB13 regulates cell death, defense and flowering time in Arabidopsis

The components in plant signal transduction pathways are intertwined and affect each other to coordinate plant growth, development and defenses to stresses. The role of ubiquitination in connecting these pathways, particularly plant innate immunity and flowering, is largely unknown. Here we report the dual roles for the Arabidopsis ( Arabidopsis thaliana ) Plant U-Box protein 13 (PUB13) in defense and flowering time control. In vitro ubiquitination assays indicated that PUB13 is an active E3 ubiquitin ligase and the intact U-box domain is required for the E3 ligase activity. Disruption of PUB13 gene by T-DNA insertion results in spontaneous cell death, accumulation of H 2 O 2 and salicylic acid (SA), and elevated resistance to biotrophic pathogens but increased susceptibility to necrotrophic pathogens. The cell death, H 2 O 2 accumulation, and resistance to necrotrophic pathogens in pub13 are enhanced when plants are pre-treated with high humidity. Importantly, pub13 also shows early flowering under middle- and long-day conditions in which the expression of SOC1 and FT is induced while FLC expression is suppressed. Finally, we found that two components involved in SA-mediated signaling pathway, SID2 and PAD4, are required for the defense and flowering time phenotypes caused by loss of function of PUB13. Taken together, our data demonstrate that PUB13 acts as an important node connecting SA-dependent defense signaling and flowering time regulation in Arabidopsis. These data indicates that signaling in the control of plant defense and plant flowering time is interconnected at WIN3. In this study, we show that the U-box type E3 ubiquitin ligase PUB13, the ortholog of rice SPL11 in Arabidopsis, regulates cell death, broad-spectrum disease resistance to

expression analysis showed that the transcription of PUB13 is completely abolished in pub13 ( Figure S1B). Under LD growth conditions [23°C, 70% relative humidity (RH), 16 hr daylight/8 hr dark], we found that the lower leaves of the pub13 mutant displayed chlorosis and lesion mimic phenotypes spontaneously, and the chlorosis was formed from leaf edge to main vein ( Figure 3A). To check whether this lesion mimic phenotype is associated with cell death, the leaves of pub13 were subjected for trypan blue staining.
The lower leaves of pub13 with chlorosis were stained blue whereas the leaves at same position in wild-type Col-0 were not ( Figure 3B, upper panel). In the middle-aged green leaves of pub13, i.e. the seventh or eighth leaf of 4-week-old plant, cell death was also detected though no macroscopic death was visible ( Figure S2B, upper panel).
To confirm whether the cell death is caused by the knockout of PUB13, we transformed the intact PUB13 CDS with 35S promoter into pub13 and a PUB13 RNAi construct into Col-0. Trypan blue staining assays showed that the non-cell death phenotype was restored in the PUB13-complemented pub13 transgenic plants ( Figure   3B, upper panel) while the lower leaves of PUB13 RNAi lines showed clear cell death same as that in pub13 ( Figure 3B, upper panel). These results demonstrate that PUB13 negatively controls cell death in Arabidopsis.
Since reactive oxygen species (ROS) are key players in the regulation of PCD, we measured the accumulation of hydrogen peroxide (H 2 O 2 ) in pub13, complemented pub13 and RNAi lines. The lower leaves of these lines grown under LD conditions were stained by 3,3'-diaminobenzidine tetrahydrochloride (DAB). In agreement with the cell death phenotypes, the H 2 O 2 content in lower leaves of pub13 and PUB13 RNAi The PCD is closely associated with SA via an unclear mechanism (Ludwig and Tenhaken, 2000;Kawai-Yamada et al., 2004). We introduced the SA-induction deficient mutant sid2-2 into pub13 by genetic cross and analyzed the role of SA for the cell death formation in pub13. Leaves of 4-week-old plants grown under LD and high humidity conditions (see details below) were collected for trypan blue staining. As shown in Figure 4A, cell death was suppressed significantly in the pub13sid2-2 double mutant compared to the clear cell death phenotype in pub13, and no visible cell death was observed in sid2-2 and Col-0. Another SA-deficient mutant, pad4, was also introduced into pub13 to confirm the pub13sid2-2 result. Similarly, the cell death in the pub13pad4 double mutant was considerably suppressed under LD and high humidity conditions ( Figure 4B).
The role of sid2-2 and pad4 for the H 2 O 2 accumulation in pub13 was also determined.
After high humidity treatment, the H 2 O 2 level in pub13 increased significantly compared to that of Col-0 but the H 2 O 2 content in sid2-2 is at similar level as Col-0 ( Figure 4C). Notably, the H 2 O 2 accumulation in the pub13sid2-2 double mutant was reduced to a background level. Similarly, introduction of pad4 into pub13 reduced the high H 2 O 2 accumulation in pub13 to the normal level under LD and high humidity conditions ( Figure 4D). The elimination of cell death and H 2 O 2 accumulation in pub13sid2-2 and pub13pad4 suggests that SA is required for the spontaneous cell death and elevated H 2 O 2 accumulation in pub13.

The pub13 Mutant Confers Enhanced Resistance against Biotrophic Pathogens
The PCD phenotype and H 2 O 2 accumulation in pub13 led us to investigate the function of PUB13 in plant defense. We first inoculated the Col-0 and pub13 plants grown under LD conditions with the virulence strain ES4326 of the biotrophic pathogen Pseudomonas syringae pv. maculicola. Three days after inoculation, the water-soaked lesions of pub13 leaves were smaller than that in Col-0 leaves ( Figure 5A, left panel).
To monitor the growth of bacteria, Psm ES4326 was infiltrated into pub13 and Col-0 plants at a low titer. The bacterial growth analysis showed that the amount of bacteria in Col-0 was more than that in pub13 ( Figure 5A, right panel), suggesting that the pub13 mutation confers elevated resistance to biotrophic bacterial pathogen Psm ES4326.
To test whether pub13 confers elevated resistance to biotrophic fungal pathogens as well, we inoculated both pub13 and Col-0 with Erysiphe cichoracearum UCSC1, which causes powdery mildew diseases in many plants. Abundant conidiophores and conidiophores peduncles spread all over the Col-0 leaves but much less on the pub13 leaves, and visible cell death appeared in the infected pub13 leaves at 7 dpi (days post inoculation; Figure 5B, left panel). To further confirm the resistance of pub13 to UCSC1, we stained the inoculated Col-0 and pub13 leaves with trypan blue at 10 dpi.
As shown in Figure 5B (right panel), a large number of hyphae and conidiophores existed in Col-0 while only a few of hyphae were observed in pub13. These results indicate that pub13 confers enhanced resistance to the biotrophic fungus strain UCSC1.
In addition to examining the resistance of pub13 against biotrophic pathogens, we also tested whether pub13 confers resistance to necrotrophic fungal pathogens Botrytis cinerea BO5-10 and Alternaria brassicicola AB. In contrast to biotrophic pathogens, no significant difference in disease symptom and pathogen growth was observed when the mutant and wild-type plants were inoculated with these two fungal pathogens under LD and and normal humidity conditions (70% RH) ( Figure S3C & 3D).

High Humidity
It is known that some lesion mimic mutants are sensitive to environmental stress conditions such as high humidity (Mosher et al., 2010). To test the effect of humidity on the occurrence of the phenotypes on pub13 plants, 4-week-old pub13 plants grown under regular conditions (approximately 70% RH) were subjected to high humidity (95% RH) treatment. After 48 hr high humidity treatment, the lesion mimic phenotype of pub13 was more severe than that in wild-type plants ( Figure S2A). Consistent with the lesion mimic phenotype, cell death in middle-aged leaves of pub13 after high humidity treatment was much stronger compared with that in untreated plants ( Figure   S2B, upper panel). In addition, the middle-aged leaves from humidity-treated pub13 but not from treated Col-0 exhibited abundant H 2 O 2 accumulation, while there was no obvious H 2 O 2 accumulation in the middle-aged leaves of untreated pub13 ( Figure S2B, lower panel). Humidity, therefore, can enhance cell death and H 2 O 2 accumulation in the pub13 plants.
To test whether the enhanced cell death and H 2 O 2 accumulation in pub13 is correlated with its elevated resistance to pathogen infections, we monitored the level of resistance of pub13 and Col-0 plants pre-treated with high humidity for 24 hr. The treated and untreated plants were inoculated with the biotrophic pathogens Pseudomonas syringae pv. tomato DC3000 and Hyaloperonospora parasitica Noco2.
Analysis of bacterial growth also showed that pub13 plants were more resistant against DC3000 than Col-0 ( Figure S3A, lower panel). Interestingly, a much larger difference of resistance against DC3000 was observed between the high humidity pretreated pub13 and Col-0 ( Figure S3A). High humidity pretreated leaves of Col-0 displayed atrophic and chlorosis phenotype, whereas only some local chlorosis existed on pub13 leaves. Similar to the inoculation with DC3000, pub13 was slightly more resistant than Col-0 to Noco2 without humidity treatment ( Figure S3B). After pre-treatment of high humidity, however, the resistance of pub13 was much increased, as revealed by less chlorosis symptoms and smaller lesions on the leaves ( Figure S3B).
Resistance of pub13 against necrotrophic pathogens Botrytis cinerea and Alternaria brassicicola after high humidity treatment was also examined. The detached leaves of pub13 and Col-0 did not show any difference in lesion size under regular humidity after inoculation with B. cinerea BO5-1 ( Figure S3C). Nevertheless, after high humidity treatment, leaves of pub13 plants displayed severe rotted symptoms and large lesions while no visible necrotic lesion occurred in leaves of Col-0 plants ( Figure S3C). Similar to challenge by B. cinerea, there was no visible symptoms difference to A. brassicicola AB between the pub13 and Col-0 under normal humidity ( Figure S3D). Conversely, pub13 was much more susceptible to A. brassicicola AB than Col-0 when the plants were pretreated with high humidity ( Figure S3D), corroborating the observation that high humidity promotes susceptibility of pub13 against necrotrophic pathogens.
PAD4 and SID2 Are Required for the PUB13-mediated Resistance to P. s. maculicola ES4326 but not to B. cinerea BO5-10 As mentioned above, the cell death and H 2 O 2 accumulation in pub13 are dependent on the SA signaling components SID2 and PAD4. To test the function of SID2 and PAD4 genes in the PUB13-mediated defense pathway, we inoculated the pub13sid2-2 and pub13pad4 plants with the biotrophic pathogen ES4326. Under regular humidity and LD conditions, pub13 displayed elevated but sid2-2 showed reduced resistance against ES4326 ( Figure 6A). As expected, the resistance of the pub13sid2-2 double mutant against ES4326 was markedly reduced to a level even more susceptible than Col-0.
Similarly, the pub13pad4 plants were a slightly more susceptible than Col-0 ( Figure   6B). In contrast to the biotrophic pathogens, the resistance level of pub13sid2-2 and pub13pad4 plants against the necrotrophic pathogen BO5-10 was comparable to that of pub13 plants after high humidity treatment ( Figure 6C). Taken together, these results demonstrated that SID2 and PAD4 are required for PUB13-mediated resistance to the biotrophic pathogen ES4326 but are dispensable for the resistance to the necrotrophic pathogen BO5-10.

The pub13 Mutant Contains Elevated Level of SA
The intimate involvement of PUB13 in host defense against biotrophic and necrotrophic pathogens prompted us to determine the expression level of known defense-related genes in pub13. Transcriptional level of PR1, a marker gene in the SA signaling pathway, and PDF1.2, a marker gene of the jasmonic acid (JA) and ethylene (ET) signaling pathway, were examined in Col-0 and pub13 plants grown under LD conditions using real time-PCR. As shown in Figure S4A, the transcriptional level of PR1 was significantly induced in pub13 compared with Col-0, whereas PDF1.2 was markedly suppressed in pub13. We then determined the SA level in the Col-0 and pub13 plants by HPLC. Total SA was extracted from 0.2 g of fresh weight leaves of 4-week-old Col-0 and pub13 plants grown under LD conditions. As expected, the SA content in pub13 was 63% higher than in Col-0 ( Figure S4B). Although the SA level was undetectable in both SA signal deficiency mutants sid2-2 and pad4, the increased SA level in pub13 was greatly reduced in the pub13sid2-2 and pub13pad4 double mutant plants, which was only 10% and 32% of that of Col-0, respectively ( Figure S4B), suggesting that PUB13 negatively regulates SA accumulation via both SID2 and PAD4.

PUB13 Is a Negative Regulator of Flowering Time
During the reproductive development stage of the pub13 mutant, we found that the mutant displayed altered flowering time. We grew the pub13 plants under different photoperiod conditions: SD (8 hr daylight/16 hr dark), MD (middle day, 12 hr daylight/12 hr dark), and LD (16 hr daylight/8 hr dark). Under SD conditions, there was no significant difference between Col-0 and pub13 in flowering time, except a few pub13 plants exhibited slightly earlier flowering than Col-0 ( Figure 7A, left panel).
However, under LD conditions, pub13 flowered about 4 days earlier than Col-0 ( Figure   7A, right panel). The early flowering phenotype of pub13 under MD conditions was even more significant compared with Col-0 ( Figure 7A, middle panel). The leaf number of pub13 before flowering was 4 and 6 less than Col-0 under LD and MD conditions, respectively, while the leaf number of pub13 was slightly less than Col-0 under SD conditions ( Figure 7B). Under LD conditions, the PUB13 RNAi plants also exhibited early flowering, however, the early flowering phenotype was abolished in the complemented pub13 plants ( Figure S5). These results indicate that PUB13 acts as a suppressor of flowering time.
To understand how PUB13 regulates floral transition, we analyzed the transcript levels of floral repressor FLC and floral activators FT and SOC1 in pub13 under SD, MD and LD conditions. As shown in Figure 7C, the transcript level of FLC in Col-0 was higher than pub13 under LD and MD conditions, and there was no visible difference under SD conditions. On the contrary, the transcriptional level of SOC1 in Col-0 was lower than in pub13 under LD, MD and SD conditions. Similarly, the transcript level of FT in Col-0 was also lower than in pub13 under LD, especially under MD conditions. Taken together, these data suggested that PUB13 regulates flowering time probably through the SOC1-mediated flowering pathway.

PUB13 Regulates Flowering Time Mainly through a SA-dependent Pathway
SA is not only a critical regulator of plant defense but also is involved in the regulation of plant flowering time (Martinez et al., 2004). Thus we investigated the relationship between the PUB13-mediated flowering time and the SA pathway. To this purpose, flowering time of pub13sid2-2 and pub13pad4 was examined under LD conditions. As shown in Figure 8A, early flowering observed in pub13 was suppressed in both pub13sid2-2 and pub13pad4 while the flowering time of sid2-2 and pad4 single mutants was the same as Col-0. The leaf number of pub13sid2-2 and pub13pad4 before the appearance of the first floral bud was almost same as Col-0, which was 4 to 5 leaves more than pub13 ( Figure 8B). Analysis of the expression level of flowering marker genes in these plants under LD conditions revealed that the expression of FLC was suppressed in pub13 but was restored in pub13sid2-2 and pub13pad4, and the expression level of FLC in sid2-2 and pad4 was comparable to that in Col-0 ( Figure 8C).
On the contrary, the transcript levels of SOC1 and FT were increased in pub13 but were similar in pad4, sid2-2, and Col-0. However, the expression level of SOC1 and FT was reduced in pub13sid2-2 and pub13pad4 compared to that in pub13 ( Figure 8C). These results suggest that PUB13 regulates flowering time in a SA signaling-dependent manner.

DISSCUSSION
The ubiquitin proteasome system (UPS)-mediated protein modification and degradation has been recognized as a critical mechanism in the regulation of numerous cellular processes in plants. The importance of the UPS in plant innate immunity and flowering has been well-documented in plants (Henriques et al., 2009;Trujillo and Shirasu, 2010). Many UPS-related components have been implicated in either of the two biological processes. Nevertheless, the interconnection between the signaling pathways underlying these two processes via the UPS has yet been reported. In this study, we extensively analyzed the functions of the Arabidopsis PUB13 gene in both innate immunity and flowering. We found that PUB13 encodes a U-box/ARM repeat protein endowed with E3 ligase activity. Genetic and physiological analysis revealed that PUB13 negatively regulates cell death, H 2 O 2 accumulation and defense against biotrophs but positively regulates the resistance to necrotrophic pathogens. We discovered that PUB13 is a negative regulator of flowering time under MD and LD conditions, and the PUB13-mediated regulation of flowering time is probably through the SOC1-mediated signaling pathway. Our results revealed dual roles for PUB13 and provided novel evidence that innate immunity and development are interconnected via the UPS in Arabidopsis.
SA is a critical signaling molecule in the pathways of the local and systemic resistance in plants. In pub13, the elevated SA level is associated with enhanced defense responses. After suppressing SA in pub13 through introduction of sid2-2 or pad4, the enhanced defense responses of pub13 are largely repressed. Therefore, PUB13 regulates plant defense responses through a SA-dependant pathway. Stresses usually can promote plant early flowering and SA accumulation (Wada et al., 2009;Wada and Takeno, 2010). Previous research showed that SA is a positive regulator of flowering not only in stressed plants but also in non-stressed plants (Martinez et al., 2004). We found that the SA level, cell death and H 2 O 2 accumulation, which are usually altered as responses to stress, are elevated in pub13 under non-stressed environment, suggesting that these responses in pub13 trigger early flowering perhaps by mimicking stress signaling. Sequence and gene structure analyses revealed that the PUB13 is the putative ortholog of the rice E3 ligase gene SPL11. SPL11 was found to negatively regulate cell death and defense but positively regulates flowering time under LD conditions probably through ubiquitination of SPIN1 in rice (Zeng et al., 2004;Vega-Sanchez et al., 2008). In this study, we report that PUB13 regulates cell death, defense, as well as flowering time through a SA-dependent pathway. Although PUB13 plays similar role to SPL11 in defense, it acts as a negative regulator of flowering time in Arabidopsis under LD and MD conditions. This difference is due to that Arabidopsis is a LD plant and rice is a SD plant. Interestingly, our genetic complementation showed that the early flowering and cell death phenotypes of pub13 were restored when the rice Spl11 gene was expressed in the pub13 plants under the control of the 35S promoter ( Figure S6), indicating the functions of PUB13/SPL11 are highly conserved in dicot and monocot plants. Further characterization of PUB13/SPL11 and other associated components, therefore, should provide exciting insights into the interconnection and coordination of innate immunity and development in plants.

Trypan blue and DAB staining
Trypan blue staining was performed for cell death assay as described previously (Bowling et al., 1994). To determine the accumulation of H 2 O 2 , we stained the selected leaves with DAB as described previously (Wohlgemuth et al., 2002). Briefly, leaves were stained with 0.1% (w/v) DAB for 8 hr in the dark, destained with 95% ethanol, and preserved in 50% ethanol. The leaves for trypan blue or DAB staining were pretreated with high humidity for 48 hr.

Pathogens inoculation
Four weeks old plants grown under LD conditions were inoculated with different pathogens. To detect the humidity effect on disease resistance, the plants were treated with high humidity (95% RH) for 24 hr, then used the treated plants or detached leaves for inoculation with different pathogens. Psm ES4326 and Pst DC3000 were sprayed on the plants with 1×10 8 CFU mL -1 for disease symptoms, or injected with 5×10 5 CFU mL -1 for bacterial growth assay. The flg22-protection analysis was performed as described previously (Zhang et al., 2010). E. cichoracearum UCSC1 was inoculated as described (Vorwerk et al., 2007), and the macroscopic symptoms were observed at 7 dpi and the microscopic symptoms were checked at 10 dpi after trypan blue staining.
For B. cinerea inoculation, the fungus was cultured on PDA plate for 2 weeks at 24°C with a 12 hr photoperiod. The fungal culture was washed by water and filtrated with a nylon mesh. Then the conidia were resuspended in potato dextrose broth and adjusted the concentration to 1×10 4 conidia mL -1 . The detached rosette leaves were placed in Petri dishes containing 0.8% agar and 5μL conidia suspension was dropped on the leaf surface. The Petri dishes with the inoculated leaves were incubated at 22°C with a 12 hr photoperiod. The diameter of lesion was measured at 3 dpi. Inoculation of detached leaves with Alternaria brassicicola (1×10 5 spores mL -1 ) was performed as described previously (van Wees et al., 2003). Pictures were taken and lesion diameter was measured at 3 dpi. For the inoculation of H. parasitica Noco2, plants were sprayed with 1×10 5 spores mL -1 of Noco2, then the inoculated plants were kept under 90% RH humidity and 17°C conditions. Symptoms were obseved and lesion diameter was measured at 7 dpi.

E3 ubiquitin ligase activity assay
The full length CDS of PUB13 (

ACKNOWLEDGMENTS
We are grateful to Drs. Xinnian Dong and Mohan Rajinikanth at Duke University for providing sid2-2 and pad4 seeds. This project is supported by grants from  An in vitro ubiquitination assay was performed with GST-PUB13/V 273 R fusion proteins in the presence or absence of wheat E1, human E2 (UBCH5b) and His-tagged ubiquitin.
The numbers on the left denote the molecular mass of marker proteins in kilodaltons.
Ubiquitination signal was detected using nickel-horseradish peroxidase (upper panel).
Expression of GST-fused proteins in the assay was detected with anti-GST immuno-blot (lower panel).         Four-week-old plants grown under LD conditions were treated with high humidity (95% RH) for 48 hr, and then cell death and H 2 O 2 accumulation were detected in the middle-aged leaves, i.e. the seventh or eighth leaves in the 4 week-old plants. A, Cell death in Col-0, pub13, sid2-2 and pub13sid2-2. B, Cell death in Col-0, pub13, pad4 and pub13pad4. C, H 2 O 2 accumulation in Col-0, pub13, sid2-2 and pub13sid2-2. D, H 2 O 2 accumulation in Col-0, pub13, pad4 and pub13pad4.   Four-week-old plants in (A) and (B) were grown under LD conditions and were injected with 1×10 5 CFU mL -1 of ES4326. C, Resistance of pub13sid2-2 and pub13pad4 double mutants to BO5-10. Detached leaves from high humidity pretreated plants grown under LD conditions were inoculated with 5 μL conidia suspension (1×10 4 conidia mL -1 ) of BO5-10, and lesion diameter was measured at 3 dpi. All the experiments were repeated at least twice with similar results. Student's t test was carried out to determine the significance of the difference. Capital letters indicate a significant difference at p < 0.01.  A, Flowering phenotypes of Col-0 and pub13 grown under SD, MD and LD conditions. B, Leaf number of Col-0 and pub13 under SD, MD and LD conditions. The leaf number was counted when the first flower bud appeared. Statistic analyses were carried out as in Figure  5. C, The expression of flowering marker genes in Col-0 and pub13 under SD, MD and LD conditions. Gene expression of FLC, SOC1 and FT in Col-0 and pub13 was determined by RT-PCR. Actin (ACT) was used as a control of loading.  A, Flowering phenotypes of pub13, pub13sid2-2, and pub13pad4 grown under LD conditions. B, Leaf number of Col-0, pub13, sid2-2, pub13sid2-2, pad4 and pub13pad4 grown under LD conditions. Leaf number for each genotype was counted once the first flower bud appeared. Statistic analyses were carried out as in Figure 5. C, Flowering marker genes transcription levels. Gene expression of FLC, SOC1 and FT in 4 week-old plants grown under LD conditions was detected with RT-PCR. Actin (ACT) was used as loading control.