Elevated Levels of MYB30 in the Phloem Accelerate Flowering in Arabidopsis through the Regulation of FLOWERING LOCUS T

In Arabidopsis thaliana, the R2R3 MYB-like transcription factor MYB30 is a positive regulator of the pathogen-induced hypersensitive response and of brassinosteroid and abscisic acid signaling. Here, we show that MYB30 expressed under the control of the strong phloem-specific SUC2 promoter accelerates flowering both in long and short days. Early flowering is mediated by elevated expression of FLOWERING LOCUS T (FT), which can be observed in the absence and presence of CONSTANS (CO), the main activator of FT. CO-independent activation by high MYB30 expression results in FT levels that remain below those observed in the wild-type plants, which show an additive CO-dependent activation. In contrast, TWIN SISTER OF FT (TSF) is repressed in plants expressing high levels of MYB30 in the phloem. In transient assays, MYB30 and CO additively increase the activity of a reporter construct driven by a 1 kb FT promoter. Acceleration of flowering by MYB30 does not require the presence of salicylic acid and is independent of FLC. Taken together, increased levels of MYB30, which was reported to be induced in response to the perception of pathogens, can accelerate flowering and MYB30 may thus be a candidate to mediate cross-talk between gene networks involved in biotic stress perception and flowering time.


Introduction
Optimal timing of the transition from vegetative to reproductive development is a critical step for the successful reproduction of flowering plants. Independent endogenous and environmental cues combine to determine flowering time through converging pathways. In Arabidopsis, many regulatory inputs are channeled into the transcriptional regulation of FLOWERING LOCUS T (FT) [1,2]. FT encodes a major component of florigen, which transmits a flowering stimulus from the leaves to the apical meristem [3,4]. Arabidopsis plants carrying mutations in the FT locus flower late in long days (LD), but almost as wild-type plants if grown in short days (SD) [5]. Under inductive conditions, FT is transcribed in the companion cells of the distal leaf veins, and FT protein is translocated with the assistance of FT-INTERACTING PRO-TEIN 1 (FTIP1) to sieve elements [6,7,8]. After the long distance movement, FT protein is unloaded to the shoot apical meristem cells, where it is proposed to migrate to the nuclei to form a complex with the bZIP transcription factor FLOWERING LOCUS D (FD) and 14-3-3 proteins [9,10,11]. The FT/FD complex reprograms the transcriptional networks in the meristem to induce the floral transition. SUPPRESSOR OF OVEREXPRES-activation of FT and directly binds to the FT 59 UTR in vivo [23]. Interestingly, GIGANTEA (GI) as an activator of CO seems to activate FT directly in mesophyll cells by binding to FT repressors and associating with the FT proximal promoter [24], and also other upstream regulators of CO such as FKF1 and CDFs have been reported to directly regulate FT based on ChIP results [25]. Similar to FT, its closest relative TWIN SISTER OF FT (TSF) plays a role as flowering integrator [26,27,28]. In seedlings, TSF is mainly expressed in the vascular tissue of leaf petioles and the hypocotyls [28]. Although tsf single mutants do not show obvious later flowering in LD or SD, a loss of tsf enhances the ft flowering phenotype in LDs and SDs and TSF transcription is also responsive to CO [26,27,28].
Plant hormones also provide an input to flowering time regulation. Under non-inductive SDs, Gibberellin (GA) plays an important role as demonstrated by the observation that the ga1 mutant, which is strongly impaired in GA biosynthesis, never flowers [29]. In LDs, overexpression of GA2OX7, which encodes a protein that decreases the amount of active GA in vascular leaf tissues delays flowering through decreased FT and TSF expression [30]. Brassinosteroids (BR) and salicylic acid (SA) are reported to promote flowering at least partially by repressing FLC [31]. Plants constitutively expressing the bacterial salicylic acid hydrolase encoding NahG gene and the SA-biosynthetic mutant salicylic acid induction deficient 2 (sid2) have been shown to flower late in LDs and SDs [32]. SIZ1, a PIAS-type SUMO E3 ligase related to yeast Siz (SAP and Miz) proteins [33], facilitates SUMO modification of FLD, which represses FLC through the autonomous pathway. Plants mutated in siz1 flower early and express reduced levels of FLC indicating that sumoylation is important for FLD's repressive function [34]. In addition, the early flowering in siz1 mutant plants is dependent on the presence of SA [34].
Plants are likely also to coordinate the transition to reproductive development with adverse environmental factors such as biotic and abiotic stresses although they may have to differentiate their response to the severity of the stress [35]. However, relatively little is known about the cross-talk and potential overlap between the gene networks regulating flowering time, pathogen defense and response to biotic stress. In a systematic screen for transcription factors that accelerate flowering if highly expressed in the phloem, we identified the transcription factor MYB30 as novel factor with a potential to regulate flowering from the phloem. MYB30 was first characterized for its role in cell death during the hypersensitive response (HR) to pathogens [36]. This R2R3 MYB-like factor is a positive regulator of HR and induces cell death through SA accumulation and the regulation of SA-responsive genes [37,38,39]. MYB30 expression is transiently upregulated during bacterial pathogen recognition [37]. In the BR signaling pathway, MYB30 interacts with BRI1-EMS-SUPPRESSOR 1 (BES1) and, as a cofactor, positively regulates BR-responsive target genes [40]. During germination, MYB30 is a target of SIZ1-mediated sumoylation, which appears to stabilize the protein possibly by protecting it from ubiquitin-mediated degradation [41,42]. Seeds carrying myb30 loss-of-function alleles are hypersensitive to ABA indicating that MYB30 acts on the crossroad of yet another hormonal pathway [41]. Taken together, MYB30 is an interesting candidate, which has the potential to mediate cross-talk between the flowering and stress pathways. In the following, we show how MYB30 could integrate into the flowering time regulatory network.

Plant Growth
For expression studies, seeds were sterilized in 75% ethanol and 100% ethanol for 5 min each and sowed on GM medium supplemented by 1% sucrose. After stratification at 4uC for 3 days, plants were grown in climate chambers at 22uC in LDs (16 hours light/8 hours dark). Time of transfer to climate chambers represents day 0. During the collection of samples, ZEITGEBER TIME (ZT) 16 was in the light and ZT24 was in the dark before the switch of light condition. For flowering time measurements, seeds were sowed on soil and stratified at 4uC for 3 days. Soil trays were transferred to LDs (16 hours light/8 hours dark) or SDs (8 hours light/16 hours dark) Percival cabinets or the green house as indicated.

Mutant and Transgenic Plants
The SUC2 prom ::MYB30 fusion was constructed by recombining the MYB30 open reading frame into a modified pGREEN vector in which the SUC2 promoter was inserted upstream of the GATEWAY recombination site. The original MYB30-pENTRY clone containing the full open reading frame was generated by the REGIA consortium [43]. The MYB30 prom ::GUS plants were generated by amplifying 3.8 kb of the upstream MYB30 regions with GATEWAY compatible primers and recombining the product into GW-GUS-pGREEN vector previously described [7]. The constructs were introduced into Agrobacterium tumefaciens GV3101 that carried the pSOUP helper plasmid. Transgenic plants were generated using the floral dip method [44]. Only lines that segregated a single locus T-DNA in the T2 generation were further analyzed. For SUC2 prom ::MYB30, two of three lines were early flowering of which one was selected for crosses and further analysis. Plants of ft-10, tsf-1, co-sail, NahG and flc-3 are all Colombia-0 ecotype and have been described (http://www. arabidopsis.org/). The myb30-1 (SALK_027644C) and myb30-2 (GK_022F04) mutants were ordered from NASC. Primers for plasmid construction and genotyping are listed in Table S1. Bombardment 30 mg gold particles were prepared for every 10 bombardments. First, 70% ethanol was used to wash the gold particles followed by three times sterile water. Finally the gold particles were suspended in 500 ml of 50% sterile glycerol.
In each bombardment, 15 mg DNA in total was used including 5 mg 35S prom ::RedLUC-pJAN, 5 mg 1.0 kb FT prom ::GreenLUC-pGREEN and 5 mg of either 35S prom ::CO, 35S prom ::MYB30 or 35S prom ::GW cassette vector. The DNA was mixed with 50 ml prepared gold beads, 50 ml 2.5 M CaCl 2 and 20 ml 0.1 M spermidine. After two washes with ethanol, the DNA-gold mix was resuspended in 50 ml 100% ethanol which was used for two independent bombardments. 5-10 mm Arabidopsis leaves were transformed by the Biolistic TM Particle Delivery System (BIO-RAD, PDS-1000/HE). After 12-24 hours incubation of the samples in constant light conditions, 1 mM luciferin was sprayed on the leaves and after one minute the emitted light was immediately measured with the help of a cooled CCD-camera adapted with optical filters to detect RedLUC and GreenLUC independently. The ratios of GreenLUC/RedLUC signals were calculated with the help of the excel macro Chroma-LUC TM Calculator version 1.0 (Promega).
Reverse Transcribed Quantitative-PCR (RT-qPCR) and Semi-quantitative PCR Total RNA was isolated using the RNeasy Mini kit according to instructions (Qiagen, Cat. no. 74104). 5 mg RNA was used for reverse transcription. After treatment with DNA-free TM DNAse I according to manufacturer's instructions (Ambion, Cat. no. AM1906), cDNA was generated at 42uC for 2 hours using Superscript II reverse transcriptase (Life Technologies, Cat. no. 18080-044). The reaction was terminated by incubation at 75uC for 10 min.
Quantitative PCR (qPCR) was performed using a real-time PCR cycler (Roche LC480). 10 ml PCR reactions included cDNA corresponding to 70 ng RNA template, 10 pmole forward primer, 10 pmole reverse primer and 5 ml 26SYBR Green qPCR buffer (Biorad, Car. no.18080-044). PCR products varied in length of 200,300 bp, and the program was composed of step 1:95uC, 3 min; step 2: [95uC 10 sec; 58uC 10 sec; 72uC, 20 sec] 40 cycles; and step 3: [95uC 3 min; 50uC 1 min; rise 0.5uC every 10 sec until up to 95uC]. Absence of primer dimers or unspecific PCR products was confirmed by melting curve analysis. PCR products were quantified against an internal standard generated by diluting a cDNA sample (5-fold dilution series from 1 to 0.008) prepared from SUC2::MYB30 expressing Col plants grown in LD. Standard deviations were calculated from three technical replicates. PP2A was used for sample normalization.

GUS Histochemical Staining
Young seedlings were incubated in 90% Acetone on ice for 30 min, rinsed with 50 mM sodium phosphate buffer and incubated for 24 hours at 37uC in GUS staining solution (0.5 mg/ml X-Gluc, 50 mM sodium phosphate buffer, 0.5 mM potassium ferrocyanide, 0.5 mM potassium ferricyanide, 0.1% Triton X-100). After incubation, samples were washed with 50 mM sodium phosphate buffer for 30 min and 70% ethanol several times until leaves turned white. The GUS staining was visualized and photographed under a stereomicroscope (Leica).

Strong Expression of MYB30 in Phloem Companion Cells Promotes Flowering under LDs and SDs in Arabidopsis
The key flowering time components of the photoperiod pathway, CO and FT, are expressed in phloem companion cells of the leaf [6,7,45]. In a flowering time screen under LD conditions we used the SUCROSE-PROTON SYMPORTER 2 (SUC2) promoter, which drives expression in all phloem companion cells, to identify novel factors from a library of 800 transcription factors (TF) that have a potential to influence flowering from the phloem [46]; [43,47,48]. Two independent SUC2 prom ::MYB30 transformed lines (#1 and #2) flowered earlier than wild type (WT) in LD growth conditions ( Figure S1). We selected line #2 for further analysis and confirmed the earlier flowering phenotype, which was even more pronounced under SDs compared to Col plants grown in the same condition ( Figure 1a). To establish that MYB30 (AT3G28910) was expressed in phloem companion cells in wild-type plants, we generated MYB30 prom ::GUS reporter lines in the Col background. Strong GUS signal was observed in vascular tissues of leaves, hypocotyl and roots, but expression was not restricted to the phloem ( Figure  S2). In our growth conditions, two independent T-DNA insertion lines of myb30 (Figure 1b), which did not express full length MYB30 transcript (Figure 1c) showed similar flowering time to WT under both LD and SD conditions showing that MYB30 is not required to promote flowering in WT plants under these conditions (Figure 1d).

Overexpression of MYB30 in the Phloem Increases FT and Reduces TSF Levels
Day-time specific expression is a key feature of the photoperiodic pathway genes CO and FT. To test whether MYB30 transcripts accumulate rhythmically, we measured steady-state mRNA levels during 24 h cycles in LDs and SDs. WT seedlings showed a diurnal rhythm of MYB30 transcripts with a peak at ZT16 and a relatively similar pattern under LDs and SDs (Figure 2a). Since transcript levels were high throughout the day despite the observed diurnal pattern, MYB30 is likely to be present throughout the day unless the protein is subjected to posttranscriptional regulation. In LDs, MYB30's highest expression overlaps with the strongest accumulation of CO protein.
Next, we tested whether expression of flowering time marker genes such as FT, TSF and CO changed in SUC2 prom ::MYB30 plants. We measured steady-state mRNA levels during 24 h cycles in LDs. FT transcripts at ZT16 were ,5 times enriched comparing SUC2 prom ::MYB30 to wild-type Col plants (Figure 2b and Figure S3a). An increase of FT expression can likely contribute to the earlier flowering time of SUC2 prom ::MYB30 under LDs. In contrast, CO mRNA did not change significantly, especially during daytime (Figure 2c and Figure S3b). This suggested that SUC2 prom ::MYB30 activated flowering through transcriptional regulation of FT that was not conveyed by changes in CO mRNA levels. We also measured the expression level of different floweringtime genes during the course of Arabidopsis development. With increasing age from one to five weeks, FT mRNA transcripts accumulated to significantly higher levels in SUC2 prom ::MYB30 plants (Figure 3a), whereas SVP and FLC mRNA did not obviously differentiate the genotypes during the first 4 weeks. Therefore, increase of FT mRNA by strong MYB30 expression in the phloem does not correlate with reduced FLC and SVP levels. At 5 weeks SVP is differentially expressed in SUC2 prom ::MYB30 and Col plants, which could be due to earlier transition into flowering that is observed in the transgenic plants. Surprisingly, TSF mRNA showed a pattern opposite to FT in that it was down-regulated in SUC2 prom ::MYB30 plants during development (Figure 3b). Thus, FT and TSF, which are both part of 'florigen' perform agonistically in SUC2 prom ::MYB30 expressing plants grown in LDs.

Ectopic Expression of MYB30 Accelerates Flowering and Impacts FT and TSF Expression in Absence and Presence of CO
In the photoperiod pathway, CO is the major activator of FT and TSF [26]. To answer whether ectopic expression of MYB30 accelerates flowering via the photoperiod pathway, we crossed SUC2 prom ::MYB30 plants to ft and co mutants. SUC2 prom ::MYB30;ft double mutants flowered even later than ft single mutants under LDs, whereas SUC2 prom ::MYB30;co flowered obviously earlier than co mutants (Figure 4a). This corroborated the observation that MYB30 acts upstream of FT and suggested that other genes are affected that caused an enhanced late flowering phenotype in the  ft-10 mutants. Since ectopic expression of MYB30 can accelerate flowering in the co mutant, it can be concluded that MYB30 does not require CO to impact flowering although plants with defective co flowered much later than wild-type plants. Accelerated flowering in SUC2 prom ::MYB30 was also observed under SDs (Figure 4b), which indicated that MYB30 can promote flowering independent of the photoperiod. SUC2 prom ::MYB30;ft flowered later than ft plants in both LD and SD growth conditions (Figure 4a, b). Considering that TSF was down-regulated in SUC2 prom ::MYB30, TSF was a candidate gene to explain an enhanced late flowering phenotype in SUC2 prom ::-MYB30;ft plants. To further test this, we generated SUC2 prom ::-MYB30;ft;tsf triple mutants, which showed a flowering time that was not significantly different to that of ft;tsf double mutants (Figure 4c and Figure S4). This confirms that altered TSF transcription likely explains the flowering time differences between ft, SUC2 prom ::MYB30;ft and SUC2 prom ::MYB30;ft;tsf plants. In sum, FT and TSF are both downstream factors of MYB30, but while FT is activated, TSF is repressed.
FT transcript levels were ,8 times higher in SUC2 prom ::-MYB30;co plants compared to co single mutants under LDs, which is similar to the ,5-fold increase observed in SUC2 prom ::MYB30 plants compared to WT (Figure 5a). Therefore, the effect of SUC2 prom ::MYB30 on FT expression is additive to that of CO, which is obviously much stronger (,40-fold at ZT16). This suggests that under LDs, although CO is the main activator of FT, MYB30 can activate FT through a parallel pathway (Figure 5a).

MYB30 Induces 1.0 kb FT Promoter Activity in Transient Bombardment Assay of Leaves
To test FT promoter activity in association with MYB30 and CO, we used transient bombardment of Arabidopsis leaves. In this assay, a 1 kb FT promoter was previously shown to be inducible by CO [7]. Co-bombardment of either CO or MYB30 alone with the 1 kb FT reporter showed that MYB30 alone could activate the expression of the reporter although the effect of CO was stronger ( Figure 6). Combining MYB30 and CO in the bombardment additively increased the fold-expression of the promoter over the control ( Figure 6). Thus, the bombardment assays confirmed that MYB30 and CO can affect FT expression in parallel pathways.

MYB30 Promotes Flowering Independently of SA Levels and FLC
MYB30 has previously been shown to be a positive regulator of pathogen defense, which is in part explained through increased accumulation of SA upon increased MYB30 expression [38]. In addition, MYB30 may directly regulate genes in the SA signaling pathway. SA was shown to be required for accelerated flowering observed in siz1 mutants [34] and the absence of SA in NahG plants was shown to delay flowering compared to Col-0 plants in LD and SD [32]. To test whether MYB30 promotes flowering through an SA pathway and dependent on FLC, we introduced the SUC2 prom ::MYB30 construct into NahG and flc3 mutant backgrounds by crossing. In our LD growth conditions, flowering was not changed significantly in the SA-deficient 35S prom ::NahG plants as compared to WT. In addition, SUC2 prom ::MYB30;35-S prom ::NahG plants flowered as early as SUC2 prom ::MYB30 plants, which suggested that MYB30 accelerated flowering independently of its reported effect on increasing SA levels (Figure 7a). Expression analysis of FT, TSF and the pathogenesis response marker gene PR1 further supported the flowering-time data and confirmed a role of MYB30 in inducing PR1 expression ( Figure  S5). FT was super-induced at ZT16 in SUC2 prom ::MYB30;35-S prom ::NahG plants confirming that SA was not required for the effect of MYB30 on FT expression ( Figure S5). As before, strong MYB30 expression in the phloem repressed the expression of TSF, a response which was also detected in the NahG background ( Figure S5).
As indicated by the unaltered expression of FLC in SUC2 prom ::-MYB30 expressing plants (Figure 3), early flowering mediated by MYB30 was not dependent on FLC. Compared to the flowering observed in both parents, SUC2 prom ::MYB30;flc3 double mutants showed an additive early flowering phenotype (Figure 7b).
In conclusion, increased expression of MYB30 in the phloem can modulate flowering by increasing the expression of FT through a pathway that acts either independent of or additively to CO. Increased FT expression by MYB30 is independent of flc and does not require the presence of SA.

Discussion
Working Model for the Effect of MYB30 on Flowering Time MYB30 can enter the flowering-time regulatory network by influencing the transcription rate of FT. As suggested by transient expression assays, the activation of FT by MYB30 requires sequences contained in the proximal 1 kb of the FT promoter. It is unclear, whether FT is a direct target of MYB30 or whether other factors mediate between the transcription factor and FT. In transgenic plants, MYB30 activates FT in absence of CO approximately 5-fold and a co-activation leads to a 5-fold higher induction of FT over CO-induced levels. This additive induction indicates that CO and MYB30 represent parallel inputs for FT induction. Although MYB30 overexpression reportedly leads to increased SA levels, which have been shown to accelerate flowering, we showed that MYB30-mediated induction of FT in the phloem does not depend on the presence of SA (Figure 7; [32,39]). Furthermore, neither SVP nor FLC transcript levels are significantly altered in SUC2 prom ::MYB30 plants at different developmental stages and we have shown that the induction of FT by MYB30 is independent of FLC (Figures 3 and 7).

Opposed Effects of MYB30 on the Expression of FT and TSF
Under LD conditions, FT and TSF are affected by MYB30 overexpression but their transcripts are regulated in opposite patterns in SUC2 prom ::MYB30 plants ( Figure 5). The accelerated flowering observed in wild-type Col and co mutant plants FT and TSF are paralogous genes in Arabidopsis and both encode for proteins that contribute to the florigen function [3,26,27,28]. In LDs, genetic defects of FT are epistatic to those of TSF and this is correlated with the significantly higher expression levels of the former in LDs. Both genes show substantial conservation in their proximal promoter sequences, but TSF lacks the distal enhancer that is important for high activation of FT by CO [7]. Nevertheless, FT and TSF are usually co-regulated and share CO as upstream activator as well as SVP and FLC as transcriptional repressors [26,28]. Based on promoter::GUS reporter plants, both genes are predominantly expressed in the phloem but FT expression is usually restricted to the distal leaf veins, whereas  TSF mainly expresses in the veins of hypocotyls and leaf petioles [28].
MYB30 is the first upstream factor that regulates FT and TSF transcript levels in an obvious opposite way, where FT is upregulated and TSF is down-regulated (Figure 3). FT and TSF have redundant and independent roles in the floral transition. For example, TSF but not FT is required for the acceleration of flowering in response to external cytokinin application and only TSF expression is increased in these conditions [49].
High FT levels may also cause TSF repression. Such a negative feed-back has been observed in a previous study where TSF was down-regulated in plants that contained an activation-tagged allele of FT. This is also supported by our observation that TSF levels were increased in the ft mutant compared to WT (Figure 5b). In contrast, FT expression was not obviously changed when TSF was activated [26,27]. Neither ft nor tsf loss-of function mutants show altered expression of their respective paralog but this could be due to the small overlap of their expression domains [28].

MYB30 as a New Component in Flowering Time Control?
Our data show that high levels of the transcription factor MYB30 in the phloem companion cells affect flowering time ( Figure 1). Based on GUS signals controlled by the upstream intergenic region of MYB30, the gene is present in the phloem in wild-type plants ( Figure S3). In addition, the diurnal expression pattern of MYB30 mRNA levels showed a peak at ZT16 in LDs, which overlaps with the moment of the strongest FT induction by CO in LDs (Figure 2; [15,50]). Since MYB30 expression is induced by biotic stress [38], it is possible that accelerated flowering observed under stress condition may be dependent on MYB30. Stress accelerated flowering has been reported in Arabidopsis as response to nutrient depletion [51], pathogen perception [52], temperature [53], UVC-stress and external SA application [32]. In Pharbitis nil and Lemna paucicostata, induction of flowering in response to nutrient depletion was abolished by the addition of aminooxyacetic acid, an effect that was reversed by adding SA externally [54,55,56]. However, we could not confirm the delayed flowering that was reported earlier for 35S prom ::NahG plants (Figure 7; [32]). In addition, we did not observe a delay in flowering in two independent myb30 loss-of-function mutants grown under normal LD and SD greenhouse conditions ( Figure 1). It is probable that multiple factors act together to cause a distinct stress-induced acceleration of flowering and as long as these factors are not entirely known, different experimental observations are expected considering variant culture conditions between laboratories.
Alternatively, a more constitutive role of MYB30-like genes could be masked by redundancy within the large MYB transcription factor family. MYB30 is part of a clade of 10 proteins, with MYB96 and MYB94 as closest relatives [57]. Similarly, regulation of FT by CCAAT-box binding NF-Y complexes, which are encoded by multi-gene families, was first shown by ectopic expression and required the generation of multiple stacked loss-offunction lines to demonstrate their role as genuine components of the photoperiod pathway [16,17,47].
In conclusion, the multi-pathway regulator MYB30 seems well positioned to connect gene networks regulating flowering, hormone signaling and stress perception, because it regulates genes involved in these various pathways and perceives signal inputs from the stress-related pathways. Further work will first have to focus on determining the conditions under which crosstalk between the signals is strongest, which will greatly facilitate uncovering whether MYB30 truly plays a biologically relevant role in flowering. Figure S1 Two independent transgenic lines of SUC2prom ::MYB30 flower early in LDs. Two transgenic lines of SUC2 prom ::MYB30 and Col-0 were grown in LDs, and their rosette and cauline leaves were counted. Statistical significance was determined using the Student's t-test (p,0.01).