UV-B radiation delays flowering time through changes in the PRC2 complex activity and miR156 levels in Arabidopsis thaliana

2.1 Plant material, growth conditions, and irradiation protocols

A. thaliana ecotype Col-0 was used for all experiments. The 35S::MIR156a, MIM156, spl9-4 and toe1-2/toe2-1 seed stocks were provided by Dr. Scott Poethig (University of Pennsylvania, USA). The clf-29 mutant (SALK_021003) was provided by Dr. Aman Husbands (Timmermans Lab, CSHL, USA). The flc-3 mutant and the FRI flc-2 line were provided by Dr. Julia Questa (Caroline Dean Lab, JIC, UK). The hyl1-2, se-1, and uvr8-mutant lines were provided by Dr. Javier Palatnik (IBR, Rosario, Argentina); Dr. Nahuel González-Schain (IBR, Rosario, Argentina); and Dr. Roman Ulm (University of Geneva, Switzerland), respectively. The msi1-5 (CS850909) seed stock was obtained from the ABRC stock centre. Plants were planted in soil, stratified for 3 days at 4 °C, and transferred to a growth chamber under long-day condition (16 hr light/8 hr dark) at 23 °C or short-day condition (8 hr light/16 hr dark) at 24 °C in a growth chamber (Adaptis A1000, Conviron). For the UV-B treatment, white light was supplemented with 2 Wm⁻² of UV-B (311 nm; Phillips narrowband TL/01 lamps); during 1 hr every day starting from Day 9 after transferring to the growth chamber until flowering, the zeitgeber time of UV-B treatments was 4 hr in both long- and short-day conditions. The UV-B intensity used is similar to solar UV-B reaching Rosario, Argentina, at noon in summer time. As a no UV-B control, plants were exposed for the same time under the same lamps covered with polyester filters that absorb UV-B at wavelengths lower than 320 nm (PE, 100 mm clear polyester plastic; Tap Plastics). UV radiation was measured using a UV-B/UV-A radiometer (UV203 AB radiometer; Macam Photometrics). Samples were collected immediately after the light treatments. The UV-B treatments were started at 9 days because at this time point, the first true leaves are emerging, and this allows for the identification of plants in the same developmental stage. At 5–7 days, cotyledons have just emerged or are expanding; thus, at these days, it is usually difficult to assess plant developmental stage (Figure 1a).

2.2 Determination of flowering time

Flowering time was counted as the number of rosette leaves at the moment of flowering or the number of days until the first flower opens, similar to previous reports (Koornneef, Hanhart, & van der Veen, 1991; Questa, Song, Geraldo, An, & Dean, 2016). Days were counted from the day of transfer to the growth chamber.

For the analysis of the msi1-5/+ mutants, 30 seeds for each treatment (control and UV-B) from self-pollination of msi1-5/+ mutants were planted. The flowering time for all of them was recorded as the number of rosette leaves at flowering and days to flowering. Afterwards, each plant was PCR genotyped using primers listed in Table S1, and only data for msi1-5/+ mutants were used for analyses.

2.3 RNA extraction and qRT-PCR

Total RNA was extracted from whole seedlings grown under long-day light conditions after 9th day of growth (immediately after the first day of UV-B treatment); the 12th day of growth (juvenile phase, before transition to reproductive growth); the 18th day of growth (adult phase); and the day of flowering of control plants (Figure 1a). RNA was isolated from 100 mg of pooled tissue using the TRIzol reagent (Invitrogen) according to manufacturer's instructions. Then, 2 μg of total RNA was incubated with RNase-free DNase I (1 unit/μL) following the protocol provided by the manufacturer (Promega), followed by reverse-transcription into first-strand cDNA using SuperScript II reverse transcriptase (Invitrogen), oligo(dT) and stem-loop oligos for miRNA156 and miR172 (Varkonyi-Gasic, Wu, Wood, Walton, & Hellens, 2007). The obtained cDNA was used as a template for quantitative PCR amplification in a Stratagene Mx3000P qPCR System (Agilent), using the intercalation dye SYBRGreen I (Invitrogen) as a fluorescent reporter and Platinum Taq polymerase (Invitrogen). Primers for each of the genes under study were designed or obtained from previous reports (Table S1). Amplifications were performed under the following conditions: 2 min of denaturation at 94 °C, 40 cycles at 94 °C for 15 s, 55 °C for 15 s, and 72°C for 30 s, followed by a 2 min extension at 72 °C. Three biological replicates and three technical replicates were performed for each sample. Melting curves for each PCR were determined by measuring the decrease of fluorescence with increasing temperature (from 65 to 98 °C). PCR products were run on a 2% (w/v) agarose gel to confirm the size of the amplification products and to verify the presence of a unique PCR product. Gene expression levels were normalized to that of Arabidopsis CALCIUM DEPENDENT PROTEIN KINASE3 (CPK3) whose expression was previously reported to remain unchanged after UV-B irradiation (Ulm et al., 2004).

2.4 Chromatin immunoprecipitation

For ChIP experiments, whole seedlings irradiated with UV-B for 4 days or kept under control conditions in the absence of UV-B were used. ChIP experiments were carried out as described in Casati et al. (2008). For each reaction, 4 μL of anti-H3 trimethylated K27 (ab6002) or anti-H3 (ab1791) antibodies was used (Abcam, Cambridge, MA). Two biological replicates from each sample were used. ChIP experiments were quantified by quantitative PCR (qPCR) in triplicates with appropriate primers (Table S1). For H3K27me3 analysis, data are represented as the ratio of H3K27me3/H3 to STM/H3.

2.5 Statistical analysis

Statistical evaluations with Student's t test or ANOVA and graphical representation of the data were performed using the SigmaPlot software package and Microsoft Excel. The means and SE are derived from independent biological samples, unless otherwise stated.

3.1 UV-B radiation delays flowering time in A. thaliana

In order to investigate whether UV-B radiation affects Arabidopsis flowering time, we exposed plants to a daily dose of UV-B light for 1 hr at an intensity of 2 W.m⁻², starting at 9 days of growth, when plants are developing the first true leaves (Figure 1a). We observed that UV-B radiation delays flowering time indicated both by an increase in the number of days before flowering (Figure 1b) and by an increase in the number of rosette leaves in UV-B irradiated plants both under long and short days (Figures 1c and S1a). Representative control and UV-B treated wild-type Arabidopsis plants grown under long- and short-day conditions are shown in Figures 1d and S1b, respectively. For subsequent analyses, all experiments were done under long-day conditions, and the parameter days before flowering were used to analyse flowering time.

3.2 Expression of FT, SOC1, and LFY is downregulated in plants exposed to UV-B radiation

Previous reports demonstrated that UV-B irradiation affects plant growth and development by changes in the expression of genes that regulate these processes (Casadevall et al., 2013; Casati, 2013; Casati & Walbot, 2003). To investigate whether the observed delay in flowering time is associated to changes in the expression of flowering regulators, we used qRT-PCR to compare gene expression levels of the floral integrators FT and SOC1 and the floral identity gene and floral integrator LFY in control and UV-B exposed plants at different time points starting at 12 days, when plants are in the juvenile phase of vegetative growth, right before the transition to reproductive growth (Klepikova, Logacheva, Dmitriev, & Penin, 2015); at 18 days, when plants are in the adult phase transitioning to flowering; and at 25 days, the first bolting day of control plants (Figure 2a). SOC1 was downregulated by UV-B light at 12 days, whereas FT and LFY were downregulated by this radiation both at 12 and 18 days. By the time control plants were bolting, SOC1 and LFY expression was similar in UV-B treated and untreated plants, whereas FT expression remained downregulated in UV-B treated plants (Figure 2a). All these genes are central integrators of the different internal and environmental signals perceived by the plant through diverse flowering pathways, promoting flowering once their expression reaches certain threshold levels (Huijser & Schmid, 2011; Moon et al., 2005). Therefore, the observed downregulation in their expression correlates with the delay in flowering time measured after UV-B exposure.

3.3 UV-B irradiation results in decreased H3K27me3 associated to flowering time genes

To gain insight into the mechanisms leading to the downregulation of the floral integrators after exposure to UV-B radiation, we analysed the effect of UV-B on the expression of selected members of the age pathway (miR156 and miR172) and the autonomous pathway (FVE and FLC), because these two pathways are likely to be important under the growing conditions used for the present study (Figure 2b,c). In addition, expression of CONSTANS (CO) was analysed as a component of the photoperiod pathway (Figure 2e). Finally, considering the regulation of flowering by the PRC2 complex occurs upstream of these flowering pathways, we also investigated if the expression of MSI1 and CLF, as members of this complex, was also affected by this radiation (Figure 2d). Taking into consideration the developmentally regulated expression of miR156 (Wang et al., 2009; Wu & Poethig, 2006), we analysed the expression for this miRNA at 9 days of growth, which is immediately after initiation of UV-B exposure, and no further than 12 days of growth, because accumulation beyond this point is negligible. A similar decrease in expression has been reported for FLC (Klepikova et al., 2015); therefore, its expression was only analysed at 12 days of growth. Figure 2b shows that there is an increase in miR156 expression in irradiated plants at both time points analysed, whereas miR172 expression is downregulated by UV-B. On the other hand, the floral repressor FLC from the autonomous pathway is upregulated by UV-B at 12 days, whereas FVE and CO levels were not affected by the treatment (Figure 2c and 2e), suggesting that these genes do not have a role in the UV-B effect observed.

Interestingly, the expression of both MSI1 and CLF was lower in plants exposed to UV-B than in control plants (Figure 2d). At 12 days, both genes are downregulated after UV-B exposure, whereas at 18 and 25 days, either CLF or MSI1 accumulates at lower levels, respectively. Because they are members of the PRC2 complex, downregulation of one of them or both at the different time points analysed (Figure 2d) could affect PRC2 complex activity. To explore the possibility that the UV-B treatment could be influencing the activity of SWINGER (SWN), the alternative methyltransferase of the PRC2 complex, we also analysed the UV-B regulation of this gene. Our results showed that SWN expression is not affected by UV-B radiation under our experimental settings (Figure S2). Therefore, we explored whether a decrease in the expression of MSI1 and CLF resulted in a differential deposition of the repressive mark H3K27me3 after UV-B exposure on the chromatin of the flowering genes known to be regulated by this complex. Thus, we performed ChIP on Arabidopsis control plants or daily exposed to UV-B and used qPCR to compare methylation levels for H3K27 associated to genes encoding two miR156 genes (MIR156A and MIR156C), FLC and FT. We analysed changes in the association of this repressive mark to different regions in these genes that have already been documented to be associated to H3K27me3 (Figure 3). There was a significant decrease in the H3K27me3 levels associated to MIR156C, FLC, and one of the sites analysed for MIR156A, after UV-B exposure (Figure 3a,b), which correlates with the observed upregulation of miR156 and FLC by this radiation. However, the levels of H3K27me3 were unchanged at FT chromatin in UV-B treated plants.

3.4 The response to UV-B irradiation of clf-29 and msi1-5 mutants supports the PRC2-mediated delay in flowering time

To gain further insight into the role of MSI1 and CLF in the delay of flowering time in UV-B irradiated plants, we exposed clf-29 and msi1-5/+ mutants to the same UV-B treatment previously described. These flowering mutants have been described before and characterized in their role in flowering time (Jiang et al., 2008; Lopez-Vernaza et al., 2012; Steinbach & Hennig, 2014). Interestingly, we observed no difference in the flowering time of msi1-5/+ heterozygous mutants that were exposed to UV-B light and the corresponding nonirradiated control (Figure 4a, Table S2), suggesting that in fact MSI1 has a central role in the regulation of flowering time under UV-B conditions. Similarly, when clf-29 mutants were grown under UV-B light, the delay in flowering time was not observed, presenting an early flowering phenotype compared to nonirradiated mutant plants (Figure 4a). Even though this result was unexpected, the number of rosette leaves in clf-29 mutants remained unchanged in both control and UV-B conditions (Table S2), suggesting that the delay in flowering time in response to UV-B radiation requires the action of MSI1 and CLF. Nonetheless, we analysed changes in expression of candidate genes that could explain the observed early flowering phenotype in irradiated mutants (Figure S3). In agreement with previous reports, FT and FLC expression is upregulated in clf-29 mutants compared with Col-0 plants (Figure S3a; Jiang et al., 2008), whereas SOC1 is downregulated in clf-29 mutants (Figure S3a). Upon UV-B treatment, both FT and SOC1 expression remained downregulated in clf-29 mutant background (Figure S3b). Interestingly, as opposed to our results in Col-0 plants, FLC expression is downregulated after UV-B exposure, suggesting that this change in expression levels could be the cause of the early flowering phenotype observed in clf-29 mutants. On the other hand, it is interesting to note that in our experiments, the flowering time under control conditions of clf-29 mutants, measured by counting days before flowering, is similar as flowering time in WT plants, in contrast to what was previously reported (Jiang et al., 2008). However, when flowering time was analysed by counting the rosette leaves before flowering, clf-29 flowered earlier than WT plants as previously reported (Table S2). Even though these two parameters usually correlate, in some rare cases, they do not (see, e.g., Xu et al., 2016). These differences may be due to experimental conditions.

To understand the consequences of the differential PRC2 complex activity after UV-B exposure, we initially investigated the role of FLC in the observed delay in flowering time. Therefore, we analysed the effect of UV-B radiation on flc-3 mutant plants. According to previous reports, flc-3 mutants flower earlier than Col-0 plants under control conditions (Michaels & Amasino, 1999; Figure 4b). However, and similarly to UV-B exposed Col-0 plants, flc-3 mutants also flowered later than the untreated control plants (Figure 4b). One of the most important regulators of FLC is FRIGIDA (FRI), responsible for FLC activation to delay flowering during long vernalization periods (Sheldon et al., 2000; Shindo et al., 2005). Considering that the fri allele is not active in the Col-0 ecotype, a FRI flc-2 line was also UV-B irradiated to test whether the FLC role in flowering time in response to UV-B light might be conditioned to the presence of FRI (Michaels & Amasino, 1999; Figure 4b). In agreement to previous reports, FRI flc-2 plants flowered at the same time as Col-0 plants, because the flc-2 mutation suppresses the late flowering phenotype determined by the introgression of the FRI allele. Similarly as observed for flc-3 and Col-0 plants, FRI flc-2 plants flowered later after the UV-B treatment (Figure 4b). Altogether, these results suggest that despite the PRC2 mediated change in H3K27me3 at the FLC locus, the observed delay in flowering time after UV-B exposure seems to be subordinated to a different pathway.

3.5 Decreased H3K27me3 on MIR156 genes delays developmental phase transitions

To investigate a different downstream pathway affected by the reduced activity of the PRC2 complex after UV-B exposure, we looked into the role of the age pathway, which regulates transitions from juvenile to adult and vegetative to reproductive phases through the sequential action of miR156 and miR172 and their corresponding targets (Wang, 2014; Wang et al., 2009;Wu et al., 2009 ; Wu & Poethig, 2006). The appearance of abaxial trichomes is one of the first adult traits detected on rosette leaves, indicating the transition from juvenile to adult phase is occurring (Wu et al., 2009; Wu & Poethig, 2006). Thus, we analysed whether UV-B radiation caused an extended juvenile phase through the higher accumulation of miR156. Interestingly, we observed that the appearance of the first abaxial trichome was delayed from Leaf 6 (6.3 ± 0.45) in control plants to Leaf 8 (8.4 ± 0.5) in UV-B irradiated plants (Figure 5a), suggesting that the juvenile phase is indeed extended as a consequence of exposure to UV-B light. Thus, we then analysed the expression of SPL9 and SPL10, two of the 11 miR156 targets, and AP2, TOE1, and TOE2, three of the six AP2-like targets of miR172 (Figure 5b; Hyun et al., 2016). SPL9 and SPL10 were downregulated in UV-B-treated plants, most likely as a consequence of the higher accumulation of mature miR156 (Figure 5b). On the other hand, miR172 targets were not affected by the treatment (Figure 5b), even when miR172 accumulated at lower levels in these plants (Figures 2b and 5b). However, the regulation of gene expression by miR172 is mostly translational in plants (Aukerman & Sakai, 2003); thus, it is possible that protein levels of these targets are still higher in UV-B exposed plants.

Next, we exposed spl9-4 and toe1-2/toe2-1 mutants to the daily dose of UV-B light previously described. Previous studies have shown that mutants in spl9 show no difference in flowering time compared with wild-type plants, whereas the double mutant toe1-2/toe2-1 presents an early flowering phenotype (Aukerman & Sakai, 2003; Schwarz, Grande, Bujdoso, Saedler, & Huijser, 2008; Xu et al., 2016). As expected, we observed these differences in flowering time when compared to Col-0 plants under control conditions in the absence of UV-B (Figure 5c,d). However, similar to Col-0 plants, a delay in flowering time by UV-B was observed in both mutant lines (Figure 5c,d). Considering SPL9 is only one out of 11 targets of miR156, and miR172 targets four other genes besides TOE1 and TOE2; functional redundancy may occur in the response to UV-B for these transcription factors. Afterwards, we exposed plants from a 35S::MIR156a overexpression line to UV-B radiation. These plants were previously shown to have an extended juvenile phase and a late flowering phenotype as a consequence of the constitutive accumulation of miR156 along development (Wu et al., 2009). Interestingly, in UV-B exposed 35S::MIR156a plants, the late flowering phenotype was reverted to a surprising early flowering phenotype compared to nonirradiated plants similar to that observed in clf-29 mutants (Figure 5d), indicating that the UV-B induced change in miR156 expression and the consequent modifications in the downstream regulatory pathway are necessary for the observed late flowering phenotype in Col-0 plants. The experiments were repeated using a MIM156 line, which overexpresses a transgene harboring miR156 binding site that sequesters mature miR156. In agreement with previous results, these plants flowered very early (Huijser & Schmid, 2011) both under control and after UV-B exposure, and similar to the results using 35S::MIR156a plants, they did not show a significant delay in flowering time by UV-B (Table S2).

Finally, in order to study whether the UV-B effect is related to changes in the biogenesis pathway of miR156 and/or miR172, we performed UV-B treatments on se-1 and hyl1-2 mutants. The delay in flowering time caused by UV-B light was also observed in both mutants (Figure S4), suggesting that the molecular mechanisms behind this developmental reprogramming do not involve changes in miRNA biogenesis.

3.6 UV-B-induced delay in flowering time depends on the UVR8 signalling pathway

Finally, to analyse whether the UVR8 photoreceptor is involved in the delay in flowering time caused by UV-B radiation, we analysed flowering time in uvr8 mutants (Figure 6a). Interestingly, uvr8 mutant plants showed a late flowering phenotype when compared with Col-0 plants when they were grown in the absence of UV-B, but these mutants did not show a delay in flowering time after UV-B exposure, demonstrating that the observed delay in flowering time requires the activity of this photoreceptor (Figure 6a). Moreover, the expression of CLF and MSI1 was unaffected by UV-B light in uvr8 mutants (Figure 6b). Similar results were observed for FLC, miR156, and SOC1 (Figure 6c–e), whereas the expression of FT was higher in irradiated plants after the first day of treatment but remained unchanged afterwards (Figure 6e). It is important to note that not only UVR8 modulates UV-B responses in plants but also multiple molecular pathways are involved in plant responses to this radiation. These nonspecific UV-B stress pathways include DNA damage signalling and defence and wound signalling pathways and MAP kinase activity, reactive oxygen species, and hormone signalling (Jenkins, 2017). Therefore, upregulation of FT by UV-B at Day 9 may be mediated by one or some of these other pathways. Altogether, these results demonstrate that UVR8 regulates the observed delay in flowering time by UV-B light in Arabidopsis.