Functional divergence of Arabidopsis REPRESSOR OF UV-B PHOTOMORPHOGENESIS 1 and 2 in repression of flowering

Abstract

Photoperiodic plants coordinate the timing of flowering with seasonal light cues, thereby optimizing their sexual reproductive success. The WD40-repeat protein REPRESSOR OF UV-B PHOTOMORPHOGENESIS 2 (RUP2) functions as a potent repressor of UV RESISTANCE LOCUS 8 (UVR8) photoreceptor-mediated UV-B induction of flowering under noninductive, short-day conditions in Arabidopsis (Arabidopsis thaliana); however, in contrast, the closely related RUP1 seems to play no major role. Here, analysis of chimeric Pro_RUP1:RUP2 and Pro_RUP2:RUP1 expression lines suggested that the distinct functions of RUP1 and RUP2 in repressing flowering are due to differences in both their coding and regulatory DNA sequences. Artificial altered expression using tissue-specific promoters indicated that RUP2 functions in repressing flowering when expressed in mesophyll and phloem companion cells, whereas RUP1 functions only when expressed in phloem companion cells. Endogenous RUP1 expression in vascular tissue was quantified as lower than that of RUP2, likely underlying the functional difference between RUP1 and RUP2 in repressing flowering. Taken together, our findings highlight the importance of phloem vasculature expression of RUP2 in repressing flowering under short days and identify a basis for the functional divergence of Arabidopsis RUP1 and RUP2 in regulating flowering time.

Introduction

Induction of flowering in angiosperms signifies the first committed step in sexual reproduction, which is tightly regulated by environmental cues and plant developmental stage. One of the key environmental cues for many plant species is day length (photoperiod), which synchronizes flowering with the changing seasons, enabling efficient sexual reproduction (Andres and Coupland 2012). Arabidopsis (Arabidopsis thaliana) is a facultative long-day plant; it flowers early under long days (LD) and, although much later, eventually also under short days (SD) (Song et al. 2015; Takagi et al. 2023). Photoperiodic flowering is initiated in response to sensing the inductive day length in leaves (Andres and Coupland 2012; Gendron and Staiger 2023), a process particularly involving the blue-light photoreceptor cryptochrome 2 (cry2) (Mockler et al. 2003; Endo et al. 2007). Inactivation of the CONSTITUTIVELY PHOTOMORPHOGENIC 1 (COP1)—SUPPRESSOR OF PHYA-105 (SPA) E3 ubiquitin ligase complex under LD in the phloem results in stabilization of the COP1 substrate and clock-regulated promoter of flowering CONSTANS (CO), a B-box family transcription factor that binds to and activates transcription of FLOWERING LOCUS T (FT), encoding the flowering hormone florigen (Putterill et al. 1995; Samach et al. 2000; Yanovsky and Kay 2002; An et al. 2004; Valverde et al. 2004; Laubinger et al. 2006; Endo et al. 2007; Jang et al. 2008; Liu et al. 2008, 2011; Turck et al. 2008; Lian et al. 2011; Ranjan et al. 2011; Andres and Coupland 2012; Sheerin et al. 2015; Song et al. 2015; Takagi et al. 2023). FT moves from phloem companion cells in the leaves to the shoot apical meristem, thereby initiating the generative phase and inducing flowering (Wigge et al. 2005; Corbesier et al. 2007; Jaeger and Wigge 2007; Mathieu et al. 2007; Turck et al. 2008; Song et al. 2015).

REPRESSOR OF UV-B PHOTOMORPHOGENESIS 1 (RUP1) and RUP2 are closely related WD40-repeat proteins that duplicated in the common ancestor of Brassicaceae and that function as partially redundant repressors of the UV-B photoreceptor UV RESISTANCE LOCUS 8 (UVR8) (Gruber et al. 2010; Rizzini et al. 2011; Heijde and Ulm 2013; Zhang et al. 2022). UVR8 is a homodimer in its ground state and monomerizes upon UV-B absorption through intrinsic tryptophans functioning as chromophores (Rizzini et al. 2011; Tilbrook et al. 2016; Li et al. 2020; Podolec et al. 2021a). Active UVR8 accumulates in nuclei and inhibits COP1 as well as several transcription factors (Kaiserli and Jenkins 2007; Yin et al. 2015, 2016; Yin and Ulm 2017; Lau et al. 2019; Liang et al. 2019; Sharma et al. 2019; Tavridou et al. 2020; Fang et al. 2022). The central light-signaling component, bZIP transcription factor ELONGATED HYPOCOTYL 5 (HY5) is stabilized through the UVR8-COP1 interaction and promotes downstream responses (Ulm et al. 2004; Brown et al. 2005; Oravecz et al. 2006; Favory et al. 2009; Stracke et al. 2010; Lau and Deng 2012; Heijde et al. 2013; Huang et al. 2013; Binkert et al. 2014). RUP1 and RUP2 interact with UVR8 and in a negative feedback loop facilitate UVR8 ground state reversion by redimerization (Gruber et al. 2010; Heijde and Ulm 2013; Wang et al. 2023).

Interestingly, RUP2 functions as a crucial repressor of flowering under SD in the presence of UV-B radiation (SD+UV) (Arongaus et al. 2018). Thus, strikingly, rup2 null mutants show a day-neutral flowering phenotype specifically under SD+UV, strongly indicating that RUP2 is key for photoperiodic flowering time regulation in Arabidopsis under natural conditions (Arongaus et al. 2018). It has been suggested that the activity of CO is suppressed by direct interaction between RUP2 and CO (Arongaus et al. 2018). On the other hand, RUP1 is apparently not involved in regulation of flowering despite sharing a high amino acid sequence similarity with RUP2, displaying similar regulatory function on UVR8 redimerization as RUP2, and showing similar potential interaction with CO as RUP2 when coexpressed in transient expression systems (Gruber et al. 2010; Wang et al. 2011; Heijde and Ulm 2013; Arongaus et al. 2018). The molecular mechanism underlying functional divergence of RUP1 and RUP2 remains poorly understood.

To investigate the functional differences between RUP1 and RUP2 in controlling photoperiodic flowering, we generated transgenic lines expressing RUP1 or RUP2 but with swapped promoters, and lines expressing RUP1 and RUP2 with tissue-specific promoters. In addition, we complemented these studies by observing endogenous expression patterns of RUP1 and RUP2 under UV-B, and dissected the flowering responses of combinatorial mutant lines to illustrate RUP functional interplay with key players in flowering regulation.

Results

The functional divergence of RUP1 and RUP2 in repressing flowering is associated with differences in their regulatory and coding sequences

To understand what underlies the functional divergence of RUP1 and RUP2 in repressing flowering in Arabidopsis (Arongaus et al. 2018), we generated chimeric constructs whereby the RUP1 and RUP2 coding sequences, along with a C-terminal 3xGlu-Glu epitope tag, were expressed under the control of RUP2 and RUP1 5ʹ and 3ʹ regulatory sequences, respectively, in the rup1 rup2 mutant background (Fig. 1A). For simplicity, the regulatory sequences are referred to as promoters and the resulting lines as Pro_RUP1:RUP2 and Pro_RUP2:RUP1 (Fig. 1A). Equivalent nonchimeric constructs were generated as controls, referred to as Pro_RUP1:RUP1 and Pro_RUP2:RUP2 (representing Glu-Glu-tagged genomic constructs, with both RUP1 and RUP2 being intronless genes) (Fig. 1A). Three independent transgenic lines per construct were selected in the rup1 rup2 mutant background, which provided a range of protein expression levels as well as corresponding lines expressing RUP1 or RUP2 at comparable levels for each promoter-coding sequence combination (Fig. 1B). All transgenic lines suppressed the seedling hypocotyl growth phenotype of rup1 rup2 under UV-B, as well as the rosette growth phenotype in plants grown on soil, indicating RUP functionality (Supplemental Fig. S1).

Figure 1.

RUP2 promoter-driven expression of RUP1 can repress flowering comparable to RUP1 or RUP2 promoter-driven expression of RUP2. A) Schematic representation of chimeric and nonchimeric (genomic) constructs used in this work. B) Immunoblot analysis of RUP1-Glu-Glu (RUP1), RUP2-Glu-Glu (RUP2), and UGPase (loading control) protein levels in 4-day-old seedlings grown in weak white light (WL-UV-B) or that supplemented with UV-B (WL + UV-B). For each construct, three independent transgenic lines were used. C and D) Quantification of flowering time under SD-UV-B (−) or SD + UV-B (+) in lines containing chimeric or genomic RUP1 and RUP2 constructs. The flowering time is represented by total leaf number (rosette and cauline leaves) (C) and days to bolting (D). Individual data points are depicted as orange circles; error bars represent the standard deviation (n = 5–10); significance was determined by a one-way ANOVA with Tukey pairwise comparison; shared letters indicate no statistically significant difference between the means (P > 0.05); data are representative of two independent experiments. Flowering in wild type (Col) is compared to rup1-1 (rup1), rup2-1 (rup2), rup1 rup2, rup1 rup2/Pro_RUP1:RUP1-Glu-Glu (Pro_RUP1:RUP1), rup1 rup2/Pro_RUP1:RUP2-Glu-Glu (Pro_RUP1:RUP2), rup1 rup2/Pro_RUP2:RUP1-Glu-Glu (Pro_RUP2:RUP1), and rup1 rup2/Pro_RUP2:RUP2-Glu-Glu (Pro_RUP2:RUP2).

Late flowering time was observed in all lines comparable to wild type under SD in the absence of UV-B (SD-UV-B), in agreement with the fact that neither RUP1 nor RUP2 represses photoperiodic flowering in the absence of UV-B (Fig. 1, C and D) (Arongaus et al. 2018). Under SD supplemented with UV-B (SD+UV-B), Pro_RUP1:RUP1 lines flowered as early as rup2 and rup1 rup2, whereas Pro_RUP2:RUP2 lines flowered as late as wild type and rup1 (Fig. 1, C and D). These data confirm the main role of RUP2 as repressor of early flowering under SD + UV-B and demonstrate that the nonchimeric, genomic transgenes in Pro_RUP1:RUP1 and Pro_RUP2:RUP2 lines mimic the respective endogenous RUP1 and RUP2 genes, confirming full functionality of the promoter and coding sequences.

Two out of three Pro_RUP2:RUP1 promoter-swap lines flowered as late as wild type in SD+UV-B (Fig. 1, C and D), indicating that RUP2 promoter-driven expression of RUP1 can replace RUP2 in repressing flowering. Interestingly, the Pro_RUP1:RUP2 promoter-swap lines also flowered late in SD+UV-B (Fig. 1, C and D), suggesting that, in contrast to RUP1, RUP2 can repress flowering even when expressed under control of the RUP1 promoter. We thus conclude that promoter and coding sequence differences between RUP1 and RUP2 contribute to their functional divergence and resulting distinct activities in repressing flowering.

RUP1 and RUP2 can repress flowering when expressed in phloem companion cells, while RUP2 is also functional in mesophyll cells

To further investigate how expression patterns of RUP1 and RUP2 influence their function, we generated transgenic lines expressing RUP1 and RUP2 in distinct tissues. For this, RUP1-YFP and RUP2-YFP were expressed in the rup1 rup2 mutant background under the control of Pro_RUP1, Pro_RUP2, and three tissue-specific promoters, namely Pro_ML1, Pro_CAB3, and Pro_SUC2 that impart expression mainly in epidermal, mesophyll, and phloem companion cells, respectively (Mitra et al. 1989; Sessions et al. 1999; Srivastava et al. 2008; Kirchenbauer et al. 2016) (Supplemental Fig. S2).

Pro_RUP1 and Pro_RUP2-driven expression of RUP1-YFP and RUP2-YFP, in both chimeric and genomic configurations, phenotypically mimicked corresponding lines with Glu-Glu epitope tags (Fig. 2). This importantly indicated that YFP fusion does not interfere with RUP1 or RUP2 function, and independently confirmed that RUP1 represses flowering only when expressed with the RUP2 promoter, whereas RUP2 represses flowering with either the RUP1 or RUP2 promoter. As expected, Pro_ML1:RUP1-YFP and Pro_CAB3:RUP1-YFP lines displayed late flowering in SD-UV-B (Fig. 2, A and C). However, early flowering was observed in these lines comparable to rup1 rup2 in SD+UV-B, suggesting that RUP1 was not able to repress flowering when expressed in epidermis or mesophyll (Fig. 2, A and C). Interestingly, however, Pro_SUC2:RUP1-YFP line #1 flowered late in SD+UV-B, similarly as in SD-UV-B, indicating that RUP1 can repress flowering in SD+UV-B when expressed in phloem companion cells (Fig. 2, A and C). Surprisingly, in contrast to line #1, the two other selected Pro_SUC2:RUP1-YFP expressing lines (#5 and #7) flowered much earlier than wild type in both SD−UV-B and SD+UV-B conditions (Fig. 2, A and C). Independent of the presently enigmatic mechanism, elevated RUP1-YFP levels in lines #5 and #7 compared to line #1 may underlie the constitutive early flowering phenotype of these two lines (Supplemental Fig. S3).

Figure 2.

RUP1 expression in phloem cells can repress flowering comparable to RUP2 expression in mesophyll or phloem cells. A, B) Representative images of 84-day-old transgenic plants expressing RUP1-YFP A) or RUP2-YFP B) in specific tissues of rup1 rup2, and grown in SD−UV-B or SD+UV-B. For each tissue-specific RUP-YFP construct, three independent transgenic lines were used. C, D) Quantification of flowering time of transgenic lines expressing RUP1 C) and RUP2 D) grown in SD−UV-B (−) or SD+UV-B (+). Flowering time is represented by total leaf number (rosette and cauline leaves; upper panels) and days to bolting (lower panels). Individual data points are depicted as orange circles; error bars represent the standard deviation [n = 8–10 for RUP1 experiment, C); n = 6–10 for RUP2 experiment, D)]; significance was determined by a one-way ANOVA with Tukey pairwise comparison; shared letters indicate no statistically significant difference between the means (P > 0.05).

Like Pro_ML1:RUP1-YFP lines, Pro_ML1:RUP2-YFP lines showed UV-B-dependent induction of flowering in SD, suggesting that RUP2 cannot repress UVR8-induced flowering when expressed in epidermal cells (Fig. 2, B and D). By contrast, all Pro_CAB3:RUP2-YFP lines and two of the Pro_SUC2:RUP2-YFP lines flowered late in both SD−UV-B and SD+UV-B (Fig. 2, B and D). A constitutive early flowering phenotype was observed for Pro_SUC2:RUP2-YFP line #23 (Fig. 2, B and D). Like Pro_SUC2:RUP1-YFP lines #5 and #7, Pro_SUC2:RUP2-YFP line #23 had the highest RUP2-YFP levels among the three independent lines (Supplemental Fig. S3). To determine the extent of this unexpected constitutive early flowering in Pro_SUC2-driven RUP1-YFP and RUP2-YFP expression lines, we analyzed additional lines in the T1 generation. One out of nine independent Pro_SUC2:RUP1-YFP T1 lines and none of 10 Pro_SUC2:RUP2-YFP T1 lines showed early flowering in SD−UV-B (Supplemental Fig. S4). On the other hand, each one of seven Pro_SUC2:RUP1-YFP T1 lines and six out of seven Pro_SUC2:RUP2-YFP T1 lines suppressed the early flowering phenotype of rup1 rup2 under SD+UV-B (Supplemental Fig. S4), further confirming that expression of both RUP1 and RUP2 in phloem companion cells can repress flowering in SD+UV-B. We thus conclude that RUP2-YFP can repress UVR8-induced early flowering in SD when expressed in mesophyll or phloem companion cells, whereas RUP1-YFP only represses flowering when expressed in phloem companion cells.

RUP2 accumulates at higher levels in the vasculature compared to RUP1 under short days with UV-B

Our data suggest that the differences between endogenous RUP1 and RUP2 in regulating flowering may be partially due to varying expression levels in vasculature and its phloem companion cells. We manually isolated petiole tissues (Kurenda and Farmer 2018) to examine endogenous RUP1 and RUP2 transcript levels in vasculature and vasculature-free tissue from wild-type plants grown in SD+UV-B. Successful petiole tissue dissection and the enrichment of vasculature were confirmed using the phloem companion cell-expressed gene ARABIDOPSIS H⁺-ATPASE 3 (AHA3) (DeWitt and Sussman 1995) (Fig. 3A). UVR8 transcript was detected in all petiole tissues, with slightly higher levels in vasculature-free tissue than in isolated vasculature (Fig. 3B). RUP2 expression in isolated vasculature was higher than in vasculature-free tissue (Fig. 3C), in sharp contrast to RUP1 (Fig. 3D), and in agreement with RNA-seq analysis of laser microdissected tissues that indeed classified RUP2 as a vasculature-specific gene (Berkowitz et al. 2021). After comparing RUP1 and RUP2 expression by reverse transcription quantitative PCR (RT-qPCR), RUP2 transcript level was estimated to be about 15 times higher than that of RUP1 in vasculature. In agreement, Pro_RUP2:GUS and Pro_RUP2:RUP2-GUS lines showed clear GUS staining in vasculature tissue, namely phloem, in contrast to Pro_RUP1:GUS and Pro_RUP1:RUP1-GUS lines (Fig. 3, E to H, and Supplemental Fig. S5, A and B). Moreover, Pro_RUP2:RUP2-GUS complemented the early flowering phenotype of rup1 rup2 in contrast to Pro_RUP1:RUP1-GUS (Supplemental Fig. S5C).

Figure 3.

RUP2 but not RUP1 expression in petioles is associated with vasculature tissues. A–D) Transcript analysis in vasculature isolated from petioles or in petioles without vasculature from wild-type Col-0 plants. AHA3 expression A); UVR8 expression B); RUP2 expression C); RUP1 expression D). Tissues were isolated from 63-day-old plants grown in SD+UV-B between ZT4 and ZT8. Individual data points from two independent biological replicates are depicted as orange circles. E–H) Transverse sections of 3-week-old wild-type (Col) plants expressing Pro_RUP1:GUSE) or Pro_RUP2:GUSF), and rup1 rup2 plants expressing Pro_RUP1:RUP1-GUSG) or Pro_RUP2:RUP2-GUSH) grown under SD+UV-B. Green arrows indicate xylem vessels, full magenta arrows indicate GUS signal in the phloem region and empty magenta arrows indicate the absence of GUS signal in the phloem region. Insets on the right of each picture show an enlarged view of the boxed area. GUS, β-glucuronidase; Scale bars: 60 µm. I) RUP1-Glu-Glu, RUP2-Glu-Glu, and UVR8 protein levels in vasculature isolated from petioles (Vasc.), petioles without vasculature (Petiole w/o vasc.) or intact petioles (Petiole) from rup1 rup2 plants expressing RUP1-Glu-Glu or RUP2-Glu-Glu under the control of the RUP1 or RUP2 promoter respectively. Anti-Glu-Glu and anti-UVR8^426–440 antibodies were used; actin levels shown as loading control. Plants were first grown in SD-UV-B for 8 weeks before being transferred to SD+UV-B for an additional 1 week. Tissues were isolated between zeitgeber time (ZT) 4 and ZT8 on Day 63 (ZT0 = lights on).

We further checked Glu-Glu epitope-tagged RUP1 and RUP2 protein levels in vasculature isolated from petioles, petioles without vasculature, and intact petioles from Pro_RUP1:RUP1-Glu-Glu and Pro_RUP2:RUP2-Glu-Glu lines grown in SD+UV-B. Consistent with the transcript data, RUP2-Glu-Glu protein accumulated at higher levels in isolated vasculature compared to that of RUP1-Glu-Glu protein (Fig. 3I). Moreover, UVR8 protein could also be detected in isolated vasculature (Fig. 3I). We conclude that RUP1 and RUP2 differ in their tissue expression pattern, with primarily RUP2 expression associated with leaf vasculature, which agrees with divergent RUP1 and RUP2 function in regulating flowering under SD+UV-B.

A constitutively monomeric UVR8 photoreceptor confers a partially early flowering phenotype under short days with UV-B

Our previous data indicated that RUP2 function in repressing flowering under SD+UV-B is associated with direct inhibitory interaction with CO and is thus distinct from RUP1 and RUP2-mediated inactivation of UVR8 signaling by facilitating UVR8 redimerization (Arongaus et al. 2018). To further test this hypothesis, we examined flowering time of uvr8-17D mutants, which endogenously express a constitutively monomeric UVR8^G101S allele (Podolec et al. 2021b). Interestingly, although the UV-B-induced photomorphogenic phenotypes of uvr8-17D are stronger than those of rup2 single mutants (Podolec et al. 2021b), uvr8-17D does not exhibit early flowering comparable to rup2 under SD+UV-B, albeit uvr8-17D does flower earlier than wild type (Fig. 4). This is likely because uvr8-17D shows enhanced UV-B signaling and thus also enhanced RUP2 levels that may directly inhibit CO activity next to more generally UVR8 signaling (Podolec et al. 2021b), counteracting the enhanced UVR8 signaling-induced flowering. In agreement, overexpression of RUP2-YFP further suppressed the early flowering phenotype of uvr8-17D (Fig. 4). In contrast, RUP2 overexpression did not repress the constitutively early flowering phenotype of cop1-4 (Supplemental Fig. S6), a mutant that is fully impaired in UVR8-mediated UV-B signaling (Oravecz et al. 2006; Favory et al. 2009), further supporting that RUP2 function in repressing flowering requires UVR8-mediated signaling (Arongaus et al. 2018). In agreement, uvr8-17D cop1-4 did not flower earlier than cop1-4 (Fig. 4). Our data provide additional evidence suggesting that RUP2 regulation of early flowering under SD+UV-B is independent of its established role as repressor of UVR8 signaling by facilitating UVR8 redimerization. Moreover, our findings suggest that the CO-repressive activity of RUP2 is specific to UVR8 pathway-activated CO.

Figure 4.

Overexpression of RUP2-YFP can repress early flowering of uvr8-17D grown in short days with UV-B. A, B) Quantification of flowering time of plants grown in SD−UV-B (−) or SD+UV-B (+). Flowering time is represented by total leaf number (rosette and cauline leaves) A) and days to bolting B). Individual data points are depicted as orange circles; error bars represent the standard deviation (n = 4–7); significance was determined by a one-way ANOVA with Tukey pairwise comparison; shared letters indicate no statistically significant difference between the means (P > 0.05).

rup2 mutation suppresses the late flowering phenotype of cry2 under long days with UV-B

We further considered whether the capacity of UVR8 to induce flowering and/or the function of RUP2 as repressor of UVR8-induced flowering are linked to the 8-h SD-UV-B irradiation conditions specifically. As Arabidopsis wild-type plants flower early under LD (Arongaus et al. 2018), we resorted to cry2 mutants that flower late under LD (Mockler et al. 2003), and tested whether LD+UV-B can induce early flowering in cry2 rup2 double mutants. As expected, cry2 and cry2 rup2 plants flowered late under LD−UV-B conditions compared to wild-type and rup2 plants (Fig. 5, A to C) (Mockler et al. 2003). However, by sharp contrast, cry2 rup2 plants flowered very early in LD+UV-B compared to cry2 plants, and indeed displayed a phenotype like rup2 plants under these conditions, flowering even slightly earlier than the wild type (Fig. 5, B and C). In agreement with the cry2 rup2 early flowering phenotype under LD+UV-B, expression of the florigen-encoding FT gene was up-regulated in cry2 rup2 compared with that in cry2 plants, specifically in LD+UV-B (Fig. 5, D and E). We thus conclude that neither the capacity of UVR8 to induce flowering nor the function of RUP2 as repressor of UVR8-induced flowering are specifically linked to short days. Rather, these functions of UVR8 and RUP2 are independent of day length and cry2 activity, and solely dependent on the presence of UV-B during the light period.

Figure 5.

Rup2 suppresses the late flowering phenotype of cry2 grown in long days with UV-B. A) Representative images of 44-day-old wild-type (Col), rup2-1 (rup2), cry2-1 (cry2), and cry2 rup2 Arabidopsis plants grown in LD (LD-UV-B) or that supplemented with UV-B (LD + UV-B). B and C) Quantification of flowering time of wild-type (Col), rup2, cry2, and cry2 rup2 plants growing in LDs with (+) or without (−) UV-B. The flowering time is represented by total leaf number (rosette and cauline leaves) (B) and days to bolting (C). Individual data points are depicted as orange circles; error bars represent the standard deviation (n = 9–11); significance was determined by a one-way ANOVA with Tukey pairwise comparison; shared letters indicate no statistically significant difference between the means (P > 0.05). D, E) Reverse transcription quantitative PCR (RT-qPCR) analysis of FT expression in 20-day-old wild-type (Col), cry2, rup2, and cry2 rup2 plants grown under LD-UV-B (D) or LD + UV-B (E) on soil. Samples were collected every 4 h; the presented results are representative of three independent experiments. ZT, zeitgeber time; (ZT0) lights on; (ZT16) lights off.

Discussion

Photoperiodic flowering is a key mechanism that ensures plant reproductive success (Takagi et al. 2023). Several visible-light photoreceptors have been implicated in the regulation of photoperiodic flowering, including phytochrome A (phyA), phyB, cry2, and FLAVIN-BINDING, KELCH REPEAT, F BOX (fkf1), which all contribute to day length–dependent regulation of the key transcription factor CO and in turn FT expression and flowering transition (Andres and Coupland 2012). Recent evidence has also revealed links between molecular players UVR8 and RUP2 and the transition to flowering in Arabidopsis (Wang et al. 2011; Hayes et al. 2014; Arongaus et al. 2018; Dotto et al. 2018; Zioutopoulou et al. 2022); however, a clear, integrated picture of all findings does not currently exist and the underlying mechanism(s) remain largely enigmatic.

RUP2 in particular plays a crucial role as repressor of UVR8-inducible flowering in SD+UV-B; in contrast to wild type, rup2 null mutants show a day-neutral flowering phenotype in the presence of UV-B (Arongaus et al. 2018). This represents a key function for Arabidopsis photoperiodism in flowering time regulation. The rup2 day-neutral flowering phenotype is UVR8 photoreceptor-dependent and strongly resembles the UV-B-independent early flowering phenotype of cop1 and spa mutants grown in SD (Laubinger et al. 2006; Jang et al. 2008; Liu et al. 2008; Arongaus et al. 2018) (see also Fig. 4 and Supplemental Fig. S6). Further, here we show that RUP2 flowering repression is distinct from cry2-related day length perception and linked only to UV-B perception and signaling.

Aside from repressing UVR8-mediated flowering in SD+UV-B, which is specific to RUP2 (Arongaus et al. 2018), RUP1 and RUP2 have largely redundant functions in the repression of UVR8-induced photomorphogenesis, albeit RUP2 has a slightly more important role based on comparisons between rup1 and rup2 null mutants (Gruber et al. 2010; Heijde and Ulm 2013; Findlay and Jenkins 2016; Liao et al. 2020; Podolec et al. 2021b). This is likely associated with higher RUP2 protein levels in the plant compared to RUP1 (Liao et al. 2020). Interestingly, RUP2-Glu-Glu levels were generally higher than RUP1-Glu-Glu in our genomic complementation lines, regardless of if expression was driven by the RUP1 or RUP2 promoter. Whether such a difference is due to differential mRNA stability, mRNA translation efficiency, or protein stability remains to be determined.

There are two main explanations for why RUP2 differs from RUP1 regarding repression of UVR8-induced flowering: first, expression patterns of their genes controlled by their unique promoters are different; second, their activities in repressing flowering determined by their unique protein coding sequences are different. Our results indicate that both the promoter and protein coding sequences of RUP1 and RUP2 contribute to the distinct functions of the proteins in repressing flowering. By analyzing lines expressing RUP1 and RUP2 under the control of tissue-specific promoters, we found that both RUP1 and RUP2 protein expressed at suitable levels in the phloem companion cells are able to suppress flowering in SD+UV-B. This suggests that RUP1, similar to RUP2, can repress flowering when expressed in the vasculature, which is consistent with the finding that, like RUP2, RUP1 can interact with CO in transient FRET-FLIM assays (Arongaus et al. 2018). In wild-type Arabidopsis plants, RUP1 does not function in repressing flowering in SD+UV-B likely because of comparatively lower expression in vasculature. A number of studies revealed that flowering time regulatory genes, including PHYA, PHYB, CRY2, CO, FT, COP1, SPA1, and GIGANTEA (GI), are expressed and encode proteins that function in the vasculature (Takada and Goto 2003; An et al. 2004; Endo et al. 2007; Winter et al. 2007; Ranjan et al. 2011; Lazaro et al. 2015; Kirchenbauer et al. 2016; Kim et al. 2018). Among these, the two phloem-expressed genes CO and FT encode central components in promoting flowering. FT expression is repressed by RUP2 in SD+UV-B, presumably because of the direct RUP2–CO interaction that inactivates CO and prevents it from activating FT transcription in the vasculature (Arongaus et al. 2018).

Interestingly, restriction of RUP2 expression to leaf mesophyll tissue via the CAB3 promoter still suppresses the early flowering phenotype of rup1 rup2 in SD+UV-B, suggesting that RUP2 may also act in mesophyll as a flowering repressor. On the other hand, RUP1 is not functional in mesophyll in regulating flowering. These results confirm that there are differences in activity between RUP1 and RUP2 in repressing flowering, and more specifically in mesophyll. There may be a RUP2-specific target in mesophyll that regulates flowering by an unknown manner, or, alternatively, minor amounts of RUP2, but not RUP1, may be mobile between tissues. There are data showing that the red-light photoreceptor phyB acts in the vascular bundle and mesophyll to repress flowering (Endo et al. 2007; Lazaro et al. 2015). At least in the vascular bundle, phyB is proposed to control CO stability via physical interaction with HIGH EXPRESSION OF OSMOTICALLY RESPONSIVE GENES1 (HOS1) and CO (Lazaro et al. 2015). Moreover, GI, for which expression is under control of the circadian clock, directly activates FT expression by binding to its promoter in both vasculature and mesophyll in a CO-independent manner (Sawa and Kay 2011). Here, we report that, like phyB and GI, RUP2 may also play a role in regulating flowering in both vasculature and mesophyll. RUP2, however, apparently regulates CO activity rather than CO stability in a UVR8-dependent manner to control FT transcription (Arongaus et al. 2018). Although RUP2 may potentially function as substrate-receptor part of a CULLIN 4-DAMAGED DNA BINDING PROTEIN 1 (CUL4-DDB1) E3 ubiquitin ligase (Ren et al. 2019), our previous work did not indicate that RUP2 affects CO protein stability (Arongaus et al. 2018).

In summary, our results provide an example of how duplicated genes have diverged in their regulatory function through differences in their regulatory and protein coding sequences. Further investigation is needed to understand the detailed molecular mechanism by which RUP2 affects CO activity in vascular tissue.

Materials and methods

Plant material and growth conditions

cop1-4, cry2-1, rup1-1, rup2-1, rup1-1 rup2-1, rup2-1/Pro_35S:RUP2, uvr8-6, and uvr8-17D are in the Arabidopsis (A. thaliana) Columbia (Col-0) accession (Deng et al. 1992; Mockler et al. 1999; Favory et al. 2009; Gruber et al. 2010; Podolec et al. 2021b). cop1-4 rup2-1, cry2-1 rup2-1, uvr8-17D cop1-4, and cop1-4 rup2-1/Pro_35S:RUP2 were generated by genetic crossing. Lines overexpressing RUP2-YFP in the uvr8-17D and rup2-1 background were generated by the floral dip method (Clough and Bent 1998) using the pB7YWG2 vector containing RUP2 (Gruber et al. 2010).

Growth conditions for flowering time experiments, transcript analysis and protein level analysis from adult tissues of plants grown under SD−/+UV and LD−/+UV were as described previously (Arongaus et al. 2018). Seeds on soil were stratified for 3 days at 4°C in the dark and transferred to GroBanks (CLF Plant Climatics) with Philips Master TL-D 58W/840 white-light fluorescent tubes (120 µmol m⁻² s⁻¹, measured with a LI-250 Light Meter, LI-COR Biosciences) supplemented or not with UV-B from Philips TL40W/01RS narrowband UV-B tubes (0.07 mW cm⁻², measured with a VLX-3W Ultraviolet Light Meter equipped with a CX-312 sensor; Vilber Lourmat), and a day/night temperature cycle of 22°C/18°C and light/dark cycle of 8 h/16 h (SD).

Growth conditions for hypocotyl measurements, transcript analysis, protein level analysis, and confocal laser scanning microscopy imaging from plate-grown seedlings are as described previously (Oravecz et al. 2006). Seeds were surface-sterilized and sown on half-strength Murashige and Skoog (MS) medium containing 1% (w/v) sucrose and 1% (w/v) agar. After 3 days of stratification at 4°C in the dark, plates were transferred to a white-light field with continuous irradiation of 3.6 µmol m⁻² s⁻¹ white light from Osram L18W/30 fluorescent tubes supplemented or not with 0.07 mW cm⁻² UV-B from Philips TL40W/01RS narrowband UV-B tubes (Oravecz et al. 2006).

Generation of Glu-Glu-tagged RUP1 and RUP2 transgenic lines

For Pro_RUP1:RUP1-Glu-Glu construct, three PCR fragments were first amplified (using primers listed in Supplemental Table S1): a fragment containing the 5ʹ regulatory sequence of RUP1 (2158 bp) and the RUP1 coding sequence; a fragment containing XhoI- and SacI-flanked sequence with a 3x Glu-Glu tag (EYMPMEEYMPMEEYMPME) followed by a stop codon and 3ʹ regulatory sequence of RUP1 (742 bp); a fragment containing pENTR221 backbone without the insert.

For Pro_RUP1:RUP2-Glu-Glu construct, four PCR fragments were first amplified (using primers listed in Supplemental Table S1): a fragment containing 5ʹ regulatory sequence of RUP1 (2,158 bp); a fragment containing RUP2 coding sequence; a fragment containing XhoI- and SacI-flanked sequence with a 3x Glu-Glu tag followed by a stop codon and 3ʹ regulatory sequence of RUP1 (742 bp); a fragment containing pENTR221 backbone.

For Pro_RUP2:RUP1-Glu-Glu construct, four PCR fragments were first amplified (using primers listed in Supplemental Table S1): a fragment containing 5ʹ regulatory sequence of RUP2 (2,415 bp); a fragment containing RUP1 coding sequence; a fragment containing XhoI- and SacI-flanked sequence with a 3x Glu-Glu tag followed by a stop codon and 3ʹ regulatory sequence of RUP2 (800 bp); a fragment containing pENTR221 backbone.

For Pro_RUP2:RUP2-Glu-Glu construct, three PCR fragments were first amplified (using primers listed in Supplemental Table S1): a fragment containing 5ʹ regulatory sequence of RUP2 (2,415 bp) and RUP2 coding sequence; a fragment containing XhoI- and SacI-flanked sequence with a 3x Glu-Glu tag followed by a stop codon and 3ʹ regulatory sequence of RUP2 (800 bp); a fragment containing pENTR221 backbone without the insert.

The flanking sequences of all fragments were designed to allow Gibson Assembly (Gibson et al. 2009). The assembled gene was transferred from pENTR221 into pGWB501 (Nakagawa et al. 2007) by Gateway cloning technology. The resulting binary vectors were transformed into rup1-1 rup2-1 by the floral dip method (Clough and Bent 1998). Homozygous lines with single-locus insertion were used in the T3 generation.

Generation of transgenic lines with YFP-tagged RUP1 and RUP2 with swapped and tissue-specific promoters

The Glu-Glu sequences were replaced by PCR-amplified YFP sequences with a linker sequence expressing GGGGSGGGGS at the N-terminus of the tag in the four Glu-Glu-tagged swapping constructs in pENTR221 generated above by restriction digestion and ligation via XhoI and SacI sites. The ML1 promoter (Pro_ML1; 3,384 bp), CAB3 promoter (Pro_CAB3; 1,537 bp), SUC2 promoter (Pro_SUC2; 2,128 bp) (An et al. 2004; Ranjan et al. 2011), and backbone fragments without promoters from newly generated Pro_RUP1:RUP1-YFP and Pro_RUP2:RUP2-YFP constructs in pENTR221 with flanking sequences that allow Gibson Assembly were PCR-amplified. After Gibson Assembly, the assembled fragments in pENTR221 were cloned into pGWB501 by Gateway technology. The resulting 10 constructs were used to transform rup1-1 rup2-1 by floral dipping (Clough and Bent 1998). Homozygous T3 generation material of single-locus insertion lines was used for experiments.

Additional Pro_SUC2:RUP-YFP-expressing transgenic lines allowing nondestructive FASTred selection and flowering time analysis in T1 generation were generated: SUC2 promoter (2,128 bp) sequence with att recombination sites was PCR-amplified and cloned into pDONR P4-P1R. RUP1 or RUP2 CDS with att recombination sites were PCR-amplified and cloned into pDONR221. Sequence expressing YFP with att recombination sites was PCR-amplified and cloned into pDONR P2R-P3. The three fragments generated by Gateway technology in pENTR vectors were cloned by MultiSite Gateway Technology (Thermo Fisher Scientific) into pFR7m34GW (Kalmbach et al. 2023), allowing FASTred selection of transformants (Shimada et al. 2010). The resulting two constructs were used to transform rup1-1 rup2-1 by floral dipping (Clough and Bent 1998). T1 seeds after selection under fluorescence stereomicroscope for RFP signal were used for subsequent experiments.

Generation of GUS transgenic lines

To generate Pro_RUP1:GUS and Pro_RUP2:GUS transgenic lines, promoter sequences of RUP1 (2,158 bp) or RUP2 (2,415 bp) were amplified by PCR with att recombination sites. PCR products were cloned via the pDONR221 vector into the pGWB533 vector (Nakagawa et al. 2007) using Gateway technology. The resulting constructs were transformed in Col wild type by floral dipping (Clough and Bent 1998).

To generate Pro_RUP1:RUP1-GUS and Pro_RUP2:RUP2-GUS lines, promoter sequence of RUP1 (1,500 bp) and RUP1 CDS or promoter sequence of RUP2 (1,680 bp) and RUP2 CDS from genomic DNA with att recombination sites were PCR-amplified and cloned via pDONR221 vector into the pMDC163 vector (Curtis and Grossniklaus 2003). The resulting constructs were used to transform rup1-1 rup2-1 by floral dipping (Clough and Bent 1998). T3 generation was used for experiments.

Immunoblot analysis

Protein extractions were performed as described previously (Arongaus et al. 2018). Samples were boiled and separated by SDS-PAGE and blotted onto PVDF or nitrocellulose membranes. Antibodies against Glu-Glu epitope tag (901802, BioLegend), RUP2^1–15 (Arongaus et al. 2018), UVR8^426–440 (Favory et al. 2009), actin (A0480, Sigma-Aldrich), UGPase (AS05086, Agrisera), and GFP (632381, Clontech) were used as primary antibodies. Horseradish peroxidase (HRP)-conjugated anti-rabbit and anti-mouse immunoglobulins (P0399 and P0447, Dako A/S) were used as secondary antibodies. Chemiluminescent signals were generated by ECL method and detected with an ImageQuant LAS 4000 mini CCD camera system (GE Healthcare).

Reverse transcription quantitative PCR

Total RNA was isolated using the RNeasy Plant Mini Kit with an on-column DNAse I treatment according to the manufacturer's instructions (Qiagen). Reverse transcription was performed by using a TaqMan Reverse Transcription Reagents kit according to the manufacturer's instructions (Thermo Fisher Scientific). Transcript levels were determined by a QuantStudio 5 Real-Time PCR system (Thermo Fisher Scientific) with 0.5 µl cDNA in a 10-µl reaction containing PowerUp SYBR Green Master Mix (Thermo Fisher Scientific) and gene-specific primers (Supplemental Table S2). The results were calculated by the ΔΔC_t method using 18S rRNA for normalization (Livak and Schmittgen 2001). For relative quantification of RUP1 and RUP2 transcript levels, primer efficiencies were integrated in the calculation. FT transcript levels were determined as previously described (Arongaus et al. 2018).

Vasculature isolation from petioles

Vasculature tissue was isolated from petioles of leaves between zeitgeber time (ZT) 4 to ZT8 from 9-week-old adult plants grown in SD according to a published protocol (Kurenda and Farmer 2018).

GUS staining and imaging

Plants were prefixed in 1% (w/v) formaldehyde in phosphate buffer (100 mm NaPO₄, pH 7.2) at 4°C for 1 h. After three washes in phosphate buffer, they were rinsed in GUS staining buffer [50 mm NaPO₄ buffer (pH 7.2), 10 mm EDTA, 1.5 mm potassium ferrocyanide, 1.5 mm potassium ferricyanide, 0.1% (v/v) Triton X-100], then briefly vacuum-infiltrated in staining solution [50 mm NaPO₄ (pH 7.2), 10 mm EDTA, 1.5 mm potassium ferrocyanide, 1.5 mm potassium ferricyanide, 0.5 mg/ml 5-bromo-4-chloro-3-indolyl-β-D-glucuronic acid (X-Glc), 0.1% (v/v) Triton X-100] before being incubated overnight at room temperature. The staining reaction was stopped by vacuum infiltration of 4% (w/v) formaldehyde + 0.1% (v/v) Triton X-100 in phosphate buffer followed by an overnight incubation at 4°C. Leaves 1 and 2 were sectioned from the plants, washed in phosphate buffer and then in distilled water, and dehydrated through a graded ethanol series [ethanol (v/v) 30%, 50%, 70%, 90%, and 100%, with incubations respectively of 30 min, 2 × 30 min, 3 × 20 min, 2 × 30 min, 2 × 30 min, and overnight at 4°C in 100% ethanol]. Technovit7100 was prepared according to the manufacturer's indications (Kulzer GmbH) by supplementing it with Hardener I, and samples were progressively infiltrated by incubations in 3:1, 1:1, and 1:3 mixes ethanol: Technovit7100 (2 h under agitation at room temperature each time), before finally incubating for 2 days at 4°C in 100% Technovit7100. Embedding was done in Technovit7100 supplemented with 1/15 Hardener II and 1/25 polyethylene glycol 400. Sectioning was performed with a Histocore AUTOCUT microtome (Leica) using disposable R35 blades. Sections of 4 µm were deposited on SuperFrost slides, mounted in Pertex and observed with a DM6B microscope (Leica) equipped with a 20× Fluotar NA 0.55 dry objective with differential interference contrast optics and a DMC5400 CMOS camera (Leica) used with binning 2 × 2. Images were processed with Fiji (Schindelin et al. 2012).

Clearing and imaging of YFP-tagged lines

Five-week-old plants expressing RUP1-YFP and RUP2-YFP under the control of tissue-specific promoters (and rup1 rup2 plants to assess autofluorescence) were fixed overnight at 4°C in phosphate buffer pH7.2 supplemented with 4% (v/v) formaldehyde and 0.1% (v/v) Triton X-100. They were then rinsed and washed 10 min in phosphate buffer saline pH 7.4, washed two times 10 min in distilled water, and incubated 10 days at room temperature in the dark in ClearSee, with a change of medium after 1 week (Kurihara et al. 2015). They were finally mounted in ClearSee between slide and coverslip, abaxial side facing up, with a spacer consisting of double-sided scotch tape and an additional coverslip. Confocal laser scanning microscopy was performed using a Leica TCS SP8 system on a DMi8 inverted microscope (Leica). Microscope configuration was as follows: objective HC PL APO CS2 20×/0.75 IMM used with water immersion; sampling speed 1,000 Hz; line averaging 2; pinhole 1.0 Airy unit at 525 nm. YFP was excited at 488 nm and its emission was collected between 516 and 564 nm using a HyD detector. All samples were examined and contrasted with identical settings, except for Pro_ML1:RUP2-YFP, Pro_CAB3:RUP2-YFP, Pro_SUC2:RUP1-YFP, and Pro_SUC2:RUP2-YFP samples that exhibited substantially stronger fluorescence signals. For each seedling the first true leaf was imaged, and z stacks were acquired through the whole leaf with voxels of 125 × 125 × 2,000 nm³. Images were all postprocessed identically with Fiji (Schindelin et al. 2012): for each stack a Gaussian blur of radius 0.6 pixel was applied, and a few z planes were maximum-projected either at the level of the epidermis, the abaxial mesophyll, or the phloem.

Statistical analysis

One-way ANOVA with Tukey pairwise comparison was performed using Minitab 19 software. Shared letters indicate no statistically significant difference between the means (P > 0.05).

Accession numbers

Sequence data from this work can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: AT5G57350 (AHA3), AT2G32950 (COP1), AT1G04400 (CRY2), AT1G65480 (FT), AT5G52250 (RUP1), AT5G23730 (RUP2), and AT5G63860 (UVR8).

Acknowledgments

We would like to thank Lothar Kalmbach for kindly sharing the pFR7m34GW destination vector allowing FastRed selection.

Author contributions

S.C. and R.U. conceived the project and designed the research. S.C. performed most of the research. A.B.A. generated the GUS reporter lines and contributed the cry2 rup2 flowering data. R.P. generated and contributed the uvr8-17D related material and contributed to uvr8-17D flowering data. C.F. and S.L. performed the imaging, and, together with E.D., analyzed the GUS reporter lines. S.C. and R.U. analyzed the data and wrote the paper. All authors reviewed and approved the manuscript for publication.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Characterization of rup1 rup2 plants expressing RUP1-Glu-Glu or RUP2-Glu-Glu under the control of the RUP1 or RUP2 promoter.

Supplemental Figure S2. Expression of RUP1-YFP and RUP2-YFP under the control of tissue-specific promoters.

Supplemental Figure S3. Protein levels in lines expressing RUP1-YFP and RUP2-YFP under the control of tissue-specific promoters.

Supplemental Figure S4. Flowering time of additional transgenic rup1 rup2 plants in the T1 generation expressing RUP1-YFP or RUP2-YFP under the control of the SUC2 promoter.

Supplemental Figure S5.RUP2 promoter-driven RUP2-GUS accumulates in phloem cells and complements the early flowering phenotype of rup1 rup2 in SD + UV-B.

Supplemental Figure S6. Overexpression of RUP2 is not able to repress early flowering of cop1-4.

Supplemental Table S1. Primers for cloning Glu-Glu-tagged RUP promoter-swap constructs.

Supplemental Table S2. Reverse transcription quantitative PCR (RT-qPCR) primer sequences.

Funding

This work was supported by the University of Geneva, the Swiss National Science Foundation (grants 31003A_175774 and 310030_207716 to R.U.), and an iGE3 PhD salary award (to R.P.).

Data availability

The data underlying this article are available in the article and in its online supplementary material.

Dive Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

• UVR8 Gramene: AT5G63860

• UVR8 Araport: AT5G63860

• AHA3 Gramene: AT5G57350

• AHA3 Araport: AT5G57350

• CRY2 Gramene: AT1G04400

• CRY2 Araport: AT1G04400

• COP1 Gramene: AT2G32950

• COP1 Araport: AT2G32950

• RUP1 Gramene: AT5G52250

• RUP1 Araport: AT5G52250

• RUP2 Gramene: AT5G23730

• RUP2 Araport: AT5G23730

• FT Gramene: AT1G65480

• FT Araport: AT1G65480

Author notes

Present address: Whitehead Institute for Biomedical Research, Cambridge, MA, USA.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/General-Instructions) is Roman Ulm ([email protected]).

Conflict of interest statement. None declared.