Background

The appropriate timing of flowering is critical for reproductive success in plants. Since the transition from the vegetative to the reproductive phases is irreversible, plants precisely coordinate endogenous developmental signals and environmental cues, such as changes in photoperiod, light quality and quantity, and temperature, to optimize the timing of flowering [1–3]. Both the endogenous and environmental signals are incorporated into flowering genetic pathways via the floral integrators, such as FLOWERING LOCUS T (FT) and SUPPRESSOR OF CONSTANS OVEREXPRESSION 1 (SOC1) [4, 5].

Photoperiod is a major environmental cue that profoundly affects floral induction [2, 3, 6]. Plants monitor photoperiodic changes to anticipate seasonal changes. CONSTANS (CO), which is a B-box zinc finger transcription factor [7], plays a central role in photoperiodic flowering by activating FT expression [8, 9]. Accordingly, CO-defective mutants and CO-overexpressing plants exhibit photoperiod-insensitive flowering phenotypes [10, 11].

The photoperiod-sensitive FT induction is mediated by the distinct accumulation peak of CO in late afternoon under long days (LDs), which is shaped by the coordinated actions of several ubiquitin/proteasome-dependent pathways [12]. A small group of E3 ubiquitin ligases and photoreceptors modulate the CO stability. The light signaling mediator CONSTITUTIVE PHOTOMORPHOGENIC 1 (COP1) degrades CO at night [13, 14]. In the light phase of the day, two opposing regulations occur through the actions of HIGH EXPRESSION OF OSMOTICALLY RESPONSIVE GENES 1 (HOS1) and FLAVIN-BINDING, KELCH REPEAT, F-BOX 1 (FKF1) E3 ubiquitin ligases. HOS1 degrades CO in the morning, and FKF1 stabilizes CO in late afternoon [15, 16]. The sequential actions of HOS1 and FKF1 contribute to the maintenance of CO accumulation at a basal level in the morning but at a high level in later afternoon, and thus Arabidopsis flowering is induced only under LDs [6]. Meanwhile, PHYTOCHROME B mediates CO degradation, but PHYTOCHROME A and CRYTOCHROME photoreceptors mediate CO stabilization [12]. It is notable that CO accumulation occurs at the specific time phase of the day under LDs, necessitating that photoperiodic flowering would be closely linked with the circadian clock [2, 3, 17].

Many clock-defective Arabidopsis mutants exhibit alterations in the photoperiodic sensitivity of flowering time, supporting the intimate linkage between the clock and photoperiodic flowering [18–22]. In addition, the circadian clock regulates the rhythmic expression of photoperiodic flowering genes, such as CO, FKF1, GIGANTEA (GI), and CYCLING DOF FACTOR 1 (CDF1) [23–26]. The clock allows the high-level expression of CO gene occurs in the light only under LDs [23]. The clock-controlled peak of FKF1 and GI expression in the LD afternoon renders FKF1-GI complex to be formed, which is crucial for CO accumulation [16]. On the other hand, the prevention of CO and FT expression by CDFs occurs in the morning through the clock function [26, 27]. It has been shown that the early flowering phenotypes of short-period plants, such as TIMING OF CAB EXPRESSION 1 (TOC1)-defective mutant (toc1) and CASEIN KINASE II BETA SUBUNIT 4 (CKB4)-overexpressing plants, are caused by the advanced expression peaks of photoperiodic flowering genes [28, 29].

While the altered flowering phenotypes of toc1 mutant and CKB4-overexpressing plants are entirely caused by clock defects, clock-independent control of photoperiodic flowering by clock components has also been proposed [30–32]. For instance, GI, which plays a role in regulating clock progression [33], activates FT expression by directly binding to its gene promoter independent of CO [30]. It has recently been reported that SENSITIVITY TO RED LIGHT REDUCED 1 (SRR1), which is required for normal oscillator function [34], regulates floral transition in a CO-independent manner [31]. Similarly, DE-ETIOLATED 1 (DET1), which is a transcriptional corepressor important for clock progression, acts as a floral repressor [32]. It is notable that GI, SRR1, and DET1 regulate flowering time independent of their roles in the clock function. In addition, they do not require CO, the central promoter of photoperiodic flowering.

LATE ELONGATED HYPOCOTYL (LHY) and its functional paralogue CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) constitute the central oscillator in Arabidopsis [20, 35, 36]. Arabidopsis mutants that are defective in LHY and CCA1 exhibit early flowering even under non-inductive conditions [20, 22, 37]. It has been shown that CCA1 regulates flowering time by binding to the SOC1 gene promoter [38]. However, it has not been explored at the molecular level how LHY regulates flowering time.

In this work, with an aim of clarifying the molecular mechanism by which LHY regulates flowering time, we performed molecular genetic and biochemical studies on LHY-defective mutants (lhy-7 and lhy-20). Notably, the expression peaks of photoperiodic flowering genes were shifted earlier in the lhy mutants. We found that the advanced expression phases of photoperiodic flowering genes are associated with the clock defects in the mutants and underlie the reduced photoperiodic sensitivity of the mutant flowering. By matching the external light/dark cycles to the endogenous circadian period, the early flowering phenotype of the mutants was rescued. Interestingly, CCA1 binds strongly to the FT gene promoter, but LHY does not exhibit such DNA-binding activity. Our data indicate that while LHY regulates the timing of flowering entirely via the circadian clock like TOC1, CCA1 would be functionally distinct from TOC1 and LHY in regulating photoperiodic flowering.

Results

Loss-of-function lhy mutants exhibit early flowering under both LDs and SDs

To investigate the functional roles of LHY in photoperiodic flowering, we examined the flowering phenotypes of LHY-defective mutants. We also aimed to clarify the molecular mechanisms by which LHY regulates flowering time: whether LHY affects flowering time entirely by modulating the circadian rhythms like TOC1 [28] or regulating the expression of specific flowering genes like GI [30] or both. In additional to the previously reported lhy-20 mutant [39], we also isolated a T-DNA insertional loss-of-function mutant, which was designated lhy-7 (Fig. 1a). The lhy-7 mutant contains a single copy of T-DNA insertion in the sixth exon of LHY gene. Gene expression analysis confirmed lack of LHY transcription in the lhy-7 mutant (Fig. 1b).

Fig. 1
figure 1

LHY-defective mutants exhibit early flowering under both LDs and SDs. a Isolation of an LHY-defective mutant (lhy-7). It was isolated from a pool of T-DNA insertional lines deposited in the Arabidopsis Biological Resource Center (Ohio State University, OH). bp, base pair. F and R, forward and reverse primers, respectively, used to examine the expression of LHY gene. b Lack of LHY gene expression in lhy-7 mutant. Gene expression was examined by RT-PCR. A tubulin gene (TUB) was used as control for constitutive expression. c Flowering phenotypes of lhy mutants under LDs. The previously reported lhy-20 mutant was also included in the assays [39]. Plants were grown until flowering in soil under LDs (16-h light and 8-h dark) (left panel). Rosette leaf numbers of 20 plants were averaged and statistically treated using Student t-test (*P < 0.01) for each plant genotype (right panel). Bars indicate standard error of the mean (SE). d Flowering phenotypes of lhy mutants under SDs. Plants were grown until flowering in soil under SDs (8-h light and 16-h dark) (left panel). Flowering times were measured as described in (c) (right panel)

Full size image

It has been reported that the loss-of-function lhy-20 mutant exhibit a shortening of the circadian period [39, 40]. We investigated the circadian period of the lhy-7 mutant by examining the rhythmic expression patterns of two representative clock output genes, COLD, CIRCADIAN RHYTHM, AND RNA BINDING 2 (CCR2) and CHLOROPHYLL A/B-BINDING PROTEIN 2 (CAB2) [41, 42]. We found that the lhy-7 mutant also exhibits advanced peak expressions of the CCR2 and CAB2 genes compared to those observed in Col-0 plants (Additional file 1), as has been observed with other LHY-defective mutants [20, 37, 39, 40].

We next examined the flowering phenotypes of the lhy-7 and lhy-20 mutants under different daylengths by counting the number of rosette leaves at bolting. Both the lhy mutants showed accelerated flowering under LDs (Fig. 1c). The early flowering phenotypes of the mutants were more prominent under SDs (Fig. 1d). The reduced photoperiodic sensitivity of the lhy flowering is similar to what has been observed with LHY-defective mutants in other ecotypes [20, 37], showing that the altered flowering phenotypes are associated with LHY function. Since the flowering phenotypes of the lhy-7 and lhy-20 mutants were similar each other, we chose the former for subsequent molecular assays.

Expression patterns of flowering genes are altered in lhy-7 mutant

To obtain insights into the molecular mechanism by which LHY regulates flowering time, we analyzed the expression patterns of flowering genes under LDs and SDs. In LD-grown plants, the expression of FT and SOC1 genes was slightly but detectably elevated in the lhy-7 mutant (Fig. 2a), which is in good agreement with the flowering phenotype of the mutant. Under SDs, the expression of FT gene, but not SOC1 gene, was markedly elevated in the lhy-7 mutant (Fig. 2b), indicating that the FT induction is the major cause of the early flowering of the mutant at least under SDs.

Fig. 2
figure 2

Expression patterns of flowering genes are altered in lhy-7 mutant. Plants were grown under either LDs or SDs for 10 days on 1/2 X Murashige and Skoog-agar plates (hereafter, referred to as MS-agar plates) before harvesting whole plant materials for total RNA extraction. Transcript levels were examined by quantitative real-time RT-PCR (qRT-PCR). Biological triplicates were averaged and statistically treated using Student t-test. Bars indicate SE. ZT, zeitgeber time. a Expression of flowering time genes under LDs. b Expression of flowering time genes under SDs

Full size image

It is known that the circadian clock is closely associated with photoperiodic flowering [1, 17, 37]. We therefore examined the expression patterns of photoperiodic flowering genes, such as CO, GI, and FKF1, in the lhy-7 mutant. The amplitude and waveform of CO transcription were not discernibly altered in the lhy-7 mutant under both LDs and SDs (Fig. 2a and b). In contrast, the waveforms of GI and FKF1 transcription appeared with advanced shifts of the peaks in the lhy-7 mutant under both photoperiod regimes. Considering that GI interacts with FKF1 to stabilize CO [16], it seems that the advanced expression phases of GI and FKF1 under SDs would lead to an elevation of the GI-FKF1 complex formation and stabilize CO, underscoring the FT induction and early flowering in the lhy-7 mutant.

We also investigated the expression patterns of flowering genes that belong to other flowering pathways, such as autonomous, thermosensory, and gibberellic acid (GA) pathways, under LDs and SDs. It was found that the expression patterns of autonomous pathway genes, such as FLOWERING LOCUS KH DOMAIN (FLK), FVE, and FCAγ, and GA pathway genes, such as SPYNDLY (SPY) and REPRESSOR OF GA1 (RGA1), were not altered in the lhy-7 mutant under both light regimes (Additional file 2). In addition, the transcription of two floral repressor genes, which play a role in temperature-responsive flowering, was not significantly affected in the mutant. While the transcript level of SHORT VEGETATIVE PHASE (SVP) gene in the mutant was comparable to that in Col-0 plants, the transcription of FLOWERING LOCUS M β (FLMβ) gene was marginally induced in the mutant (Additional file 2). Considering the floral repressive activity of FLMβ [43], it is apparent that the early flowering phenotype of the lhy-7 mutant is not associated with the slight induction of FLMβ gene.

FLC gene is not related with lhy-7 flowering

It has been reported that induction of the floral repressor FLOWERING LOCUS C (FLC) is linked with the late flowering phenotype of cca1 lhy double mutant grown under continuous light conditions [44]. The FLC gene is also up-regulated in the parental lhy mutant in Landsberg erecta (Ler) background. We found that FLC expression is slightly increased in the lhy-7 mutant under LDs (Fig. 2a). In contrast, the FLC expression was not discernibly elevated in the mutant under SDs (Fig. 2b), suggesting that FLC gene is not associated with the mutant flowering phenotype.

To further examine whether FLC gene is associated with the flowering phenotype of the lhy-7 mutant, we crossed the lhy-7 mutant with the FLC-defective flc-3 mutant that exhibits early flowering, more prominently under SDs [45]. We compared the flowering time of the resultant lhy-7 flc-3 double mutant with those of the parental mutants. The flowering phenotype of the lhy-7 flc-3 double mutant was not discernibly different from those of the single mutants under LDs (Fig. 3a). In contrast, the lhy-7 flc-3 double mutant flowered earlier than the single mutants under SDs (Fig. 3b), indicating that FLC gene is not directly associated with the flowering phenotype of the lhy-7 mutant.

Fig. 3
figure 3

FLC gene is not associated with the flowering phenotype of lhy-7 mutant. The lhy-7 mutant was crossed with the flc-3 mutant, resulting in lhy-7 flc-3 double mutant. Plants were grown until flowering in soil under either LDs or SDs (left panel). Rosette leaf numbers of 20 plants were averaged and statistically treated using Student t-test (*P < 0.01) (right panel). Bars indicate SE. a Flowering phenotype of lhy-7 flc-3 double mutant under LDs. b Flowering phenotype of lhy-7 flc-3 double mutant under SDs

Full size image

LHY does not bind to FT promoter

LHY and CCA1 regulate a variety of genes by directly binding to the gene promoters [35, 38, 46]. For example, the CCA1 transcription factor represses SOC1 expression by binding directly to the gene promoter [38]. We found that FT gene is significantly up-regulated in the lhy-7 mutant (Fig. 2). It was therefore suspected that LHY might repress FT expression perhaps by binding to the gene promoter.

Nucleotide sequence analysis identified a putative CCA1-binding sequence (CBS) in the FT gene promoter and a potential evening element (EE) in the first intron (Fig. 4a). To examine whether LHY binds to the CBS and EE sequences, we employed chromatin immunoprecipitation (ChIP) assays using transgenic plants overexpressing a LHY-MYC gene fusion, in which a MYC-coding sequence was fused in-frame to the 3′ end of the LHY-coding sequence (Additional file 3). We also included the transgenic plants overexpressing a MYC-CCA1 gene fusion in the assays. Both the 35S:LHY-MYC and 35:MYC-CCA1 transgenic exhibited elongated hypocotyls, disruption of circadian rhythms, and suppression of FT expression (Additional file 4), as reported previously [19, 47], confirming that the transgenic plants are relevant for ChIP assays. Quantitative ChIP-PCR runs revealed that CCA1 binds to the CBS and EE sequence elements (Fig. 4b). In contrast, LHY did not bind to the sequence elements, while it efficiently bound to the TOC1 gene promoter containing EE [35].

Fig. 4
figure 4

LHY does not bind to FT promoter. a Genomic structure of FT gene. (Left panel) Gray box indicates the gene promoter region. Black boxes indicate exons, and white boxes indicate untranslated regions. CBS, CCA1-binding sequence. EE, evening element. NB, non-binding sequence. (Right panel) CBS and EE sequences are listed. b ChIP assays on binding of CCA1 and LHY to FT promoter. A MYC-coding sequence was fused in-frame to the 5′ end of the CCA1-coding sequence and the 3′ end of the LHY-coding sequence, and the gene fusions were overexpressed driven by the Cauliflower Mosaic Virus (CaMV) 35S promoter in Col-0 plants, resulting in 35S:MYC-CCA1 and 35S:LHY-MYC, respectively. Chromatins were prepared from 7-day-old whole plants grown on MS-agar plates and immunoprecipitated using an anti-MYC antibody. Fragmented genomic DNA was eluted from the protein-DNA complexes and subjected to quantitative PCR. Biological triplicates were averaged and statistically treated using Student t-test (*P < 0.01). Bars indicate SE. The promoter sequences of eIF4a and TUB genes were included as negative controls in the assays. The promoter sequence of TIMING OF CAB EXPRESSION 1 (TOC1) gene containing EE was included as positive control [35]. c Suppression of FT transcription by CCA1. The reporter and effector constructs are illustrated (left panel). Transient GUS expression assays were performed using Arabidopsis protoplasts (right panel). Five measurements were averaged and statistically treated (t-test, *P < 0.01). Bars indicate SE

Full size image

To verify the binding of CCA1 to FT chromatin, we performed electrophoretic mobility shift assay (EMSA) using a recombinant maltose binding protein (MBP)-CCA1 fusion protein. Consistent with the ChIP data, it was found that CCA1 binds specifically to CBS and EE sequence elements (Additional files 5 and 6).

To examine whether CCA1 binding to FT chromatin influences FT expression, we performed transient β-glucuronidase (GUS) expression assays by coexpressing the FT promoter-driven GUS reporter plasmid (FT pro:GUS) and the effector plasmids (35S:CCA1 or 35S:LHY) in Arabidopsis protoplasts. The assays showed that CCA1 negatively regulates GUS expression, but LHY does not affect GUS expression (Fig. 4c), consistent with the ChIP data. The repressive effects of CCA1 on GUS expression disappeared when a reporter plasmid harboring mutations in CBS was coexpressed, indicating that the binding of CCA1 to CBS is essential for the CCA1-mediated repression of FT expression. These observations indicate that although LHY and CCA1 are known to be functionally redundant [20, 48], LHY is distinct from CCA1 in regulating FT expression.

It has been reported that GI is associated with the flowering phenotype of cca1 lhy double mutant [22]. While CCA1 is directly associated with GI promoter [38, 49], it is unknown whether LHY binds to GI promoter. ChIP assays using transgenic plants overexpressing a LHY-MYC gene fusion showed that LHY also binds to GI promoter (Additional file 7), like CCA1. Binding of both LHY and CCA1 to GI promoter is in harmony with the repression of GI expression in both CCA1- and LHY-inducible lines [50].

Clock defects underlie the reduced photoperiodic sensitivity of lhy flowering

LHY is a key component of the central oscillator of plant circadian clock. LHY-defective mutants exhibit a shortened circadian rhythm of approximately 20 h compared to that of Col-0 plants ([39], this work). We found that lhy-7 mutant exhibits early flowering with a reduced sensitivity to photoperiod. Therefore, a critical question was whether the reduced photoperiodic sensitivity of the lhy-7 flowering is interconnected with the clock defects in the mutant.

To address the question, we measured the flowering times of lhy-7 and lhy-20 mutants under light/dark (L/D) cycles that were matched to the endogenous circadian periods of the mutants. If the altered flowering times of the lhy mutants are entirely due to the clock defects, the flowering times would be restored by matching the external L/D cycles to the endogenous circadian period, as has been observed with TOC1-defective mutants [21, 28]. As expected, we found that the early flowering of the lhy mutants was completely annulled under SDs of 20 h (6.7 L: 13.3D) (Fig. 5a), which matches to the endogenous period of the mutants ([39], this work), indicating that LHY regulates flowering time entirely via the circadian clock.

Fig. 5
figure 5

Clock defects underlie the reduced photoperiodic sensitivity of lhy flowering. a Flowering times of lhy mutants under SDs of differential total duration. Plants were grown until flowering under SDs of 24-h or 20-h total duration. Those grown under SDs of 20-h total duration were photographed (left panel). Rosette leaf numbers of 20 plants were averaged and statistically treated (t-test, *P < 0.01) (right panel). Bars indicate SE. b Expression profiles of FT and clock output genes in lhy-7 mutant grown under SDs of 24-h or 20-h total duration. Transcript levels were examined by qRT-PCR. L/D, light/dark. c Flowering phenotype of lhy-7 co-101 double mutant under LDs. The lhy-7 mutant was crossed with the co-101 mutant, resulting in lhy-7 co-101 double mutant. Plants were grown until flowering in soil under LDs (left panel). Flowering times were measured as described in (a) (right panel). d Expression of FT gene in lhy-7 co-101 double mutant grown under either LDs or SDs. Transcript levels were examined by qRT-PCR. Biological triplicates were averaged. Bars indicate SE

Full size image

We also analyzed the expression profiles of flowering time genes in the lhy-7 mutant under SDs of adjusted L/D cycles. The elevation of FT transcript levels was not observed when the lhy-7 mutant was grown under SDs of 20 h (Fig. 5b and Additional file 8), as has been observed with short-period plants, such as toc1 null mutants [28, 29]. It is therefore evident that the reduced photoperiod sensitivity of the lhy-7 flowering is caused by the clock defects.

It was found that CO transcription was elevated at night under the assay conditions (Fig. 5b). Since CO protein is degraded under dark conditions [13, 14], it is unlikely that the CO induction at night is physiologically important for the flowering phenotype of the lhy-7 mutant. The waveform of GI transcription under the adjusted L/D cycles was comparable to that in Col-0 plants under SDs of 24 h. In addition, gene expression assays revealed that the expression of flowering genes, such as FLK, FVE, FCAγ, SVP, SPY, and RGA1, was not altered in the lhy-7 mutant under both SDs of 20 and 24 h. The slight induction of FLMβ gene in the lhy-7 mutant under SDs of 24 h was also observed under SDs of 20 h (Additional files 2 and 9). On the basis of the flowering times and expression patterns of flowering genes in the lhy mutants under adjusted L/D cycles, we concluded that LHY regulates photoperiodic flowering via the clock function.

To verify that LHY regulates flowering time through the clock-controlled CO-FT pathway, we generated lhy-7 co-101 double mutant by crossing the lhy-7 mutant with the CO-defective co-101 mutant. Acceleration of flowering by the lhy mutation completely disappeared in the lhy-7 co-101 double mutant under LDs (Fig. 5c). Accordingly, FT expression was slightly induced in the lhy-7 mutant, but the induction was compromised in the lhy-7 co-101 double mutant (Fig. 5d). We also examined the level of FT transcripts in the single and double mutants under SDs, since the early flowering phenotype of the single mutant was more prominent under this light regime (Fig. 1d). We found that the elevation of FT expression in the lhy-7 mutant was completely annulled in the lhy-7 co-101 double mutant (Fig. 5d). Together, these observations indicate that LHY-mediated regulation of photoperiodic flowering depends on CO function.

Discussion

LHY-mediated clock control of photoperiodic flowering

Plants sense photoperiodic changes by integrating light signals perceived by the photoreceptors and timing information provided by the circadian clock. In Arabidopsis, photoperiod-sensitive induction of the floral integrator FT is a crucial molecular event in photoperiodic flowering [6]. It is known that the coordinated action of light signals and timing information allows the CO transcription factor to accumulate specifically in late afternoon under LDs, which is a prerequisite for the LD-specific induction of FT gene [1–3, 12]. The circadian clock regulates CO activity at both the transcriptional and posttranscriptional levels [12, 16, 23]. At the transcriptional level, the clock shapes the rhythmic expression patterns of CO in a way that a high level of CO transcripts accumulates during the light phase under LDs [23]. At the posttranscriptional level, the circadian clock contributes to the stabilization of CO in late afternoon under LDs by modulating the expression of GI and FKF1 genes [16, 17].

In this study, we demonstrated that LHY, which is a core component of the central oscillator in Arabidopsis, regulates photoperiodic flowering by adjusting the rhythmic expression patterns of photoperiodic flowering genes, such as GI and FKF1. We found that the expression peaks of GI and FKF1 genes are shifted earlier in the lhy-7 mutant, which is consistent with the shortened circadian period of the mutant. A plausible explanation is that the advanced phases of GI and FKF1 expression in the lhy-7 mutant would lead to an increase in the formation of GI-FKF1 complexes during the light phase, resulting in a higher-level accumulation of CO in the mutant. In support of this view, the early flowering phenotype of the mutant was completely annulled by matching the external L/D cycles to the endogenous circadian period. Under these assay conditions, the phase shift of GI transcription was restored and the up-regulated expression of FT was suppressed to a basal level in the lhy-7 mutant. Together with the previous reports on short-period plants [28, 29], it seems that the circadian clock components, including LHY, regulates photoperiodic flowering by adjusting the expression timing of photoperiodic flowering genes.

Common and distinct roles of LHY and CCA1 in photoperiodic flowering

LHY and CCA1 are MYB motif-containing transcription factors that function at least in part redundantly in the circadian clock [19, 20, 48]. Whereas the gain-of-function mutations of both LHY and CCA1 genes disrupt circadian rhythms [19, 20], both the LHY- and CCA1-defective mutants exhibit shortened circadian periods [20, 37, 51]. The cca1 lhy double mutants show shorter circadian periods than the cca1 or lhy single mutants [20, 37, 48].

On the other hand, there have been some reports supporting distinct roles of CCA1 and LHY. For instance, LHY overexpression does not enhance pathogen resistance, whereas CCA1 overexpression induces pathogen resistance [52, 53]. Another example is the differential regulation of CCA1 and LHY expression by CCA1 HIKING EXPEDITION (CHE) and BROTHER OF LUX ARRHYTHMON (BOA), the components of the Arabidopsis circadian clock. While CHE and BOA bind directly to CCA1 gene promoter, they are not associated with LHY gene promoter [46, 54].

We found that CCA1, but not LHY, binds to FT gene promoter to repress its expression. It has been reported that CCA1 represses SOC1 expression by binding to the gene promoter [38]. We observed that SOC1 expression is not discernibly affected by lhy mutations, suggesting that LHY is not related with SOC1 expression. It is known that LHY and CCA1 form both homodimers and heterodimers in vivo [47, 55]. One possibility is that whereas common roles of the two transcription factors would be related with the LHY-CCA1 heterodimers, their distinct roles would be exerted through the homodimers.

Clock-independent control of photoperiodic flowering

It is now apparent that most clock components, including LHY, TOC1, and CKB4, regulate photoperiodic flowering via the clock function [28, 29]. Notably, the shorter the circadian periods of clock mutants, the earlier the flowering times of the mutants in most cases [37], supporting that the clock regulates photoperiodic flowering by modulating the rhythmic expression of photoperiodic flowering genes [1, 6, 17].

However, there are recent reports supporting that some clock components affect flowering time through clock-independent pathways. CCA1 regulates the expression of floral integrators by directly binding to the gene promoters ([38], this work). GI also binds directly to FT gene promoter [30]. In addition, the early flowering phenotype of DET1-defective mutants is not restored by matching the external L/D cycles to the endogenous circadian period of the mutants, indicating that DET1 negatively regulates flowering independent of its role in the clock function [32]. Together, it is likely that clock components regulate photoperiodic flowering through both CO-mediated, clock-dependent and CO-free, clock-independent pathways. It will be interesting to investigate how individual clock components regulate flowering time and how the clock-dependent and clock-independent pathways are functionally inter-connected with each other in photoperiodic flowering.

Conclusions

We investigated how LHY regulates photoperiodic flowering by performing molecular genetic and biochemical studies. LHY regulates photoperiodic flowering entirely via the circadian clock. In the LHY-defective mutants, the early flowering phenotypes and the shifted phases of photoperiodic flowering gene expression were recovered by matching the external L/D cycles to the endogenous circadian periods of the mutants. It is notable that the mechanism by which LHY regulates photoperiodic flowering is somewhat distinct from that exerted by CCA1. Our findings would contribute to better understanding of how the clock function is associated with flowering time control in response to photoperiodic signals.

Methods

Plant genotypes and growth conditions

Arabidopsis thaliana lines used were in the Columbia (Col-0) background, unless specified otherwise. Arabidopsis plants were grown either in soil or on 1/2 X Murashige and Skoog (MS)-agar plates (hereafter referred to as MS-agar plates) at 23 °C under either LDs (16-h light and 8-h dark) or SDs of 24 h (8 L: 16D) or 20 h (6.7 L: 13.3D) total duration. White light illumination (120 μmol photons m−2s−1) was provided by fluorescent FLR40D/A tubes (Osram, Seoul, Korea).

T-DNA insertional gene knockout mutants lhy-20, flc-3, and co-101 have been described previously [39, 45, 56]. The LHY-deficient lhy-7 mutant (SALK-149287) was isolated from a T-DNA insertional mutant pool deposited in the Arabidopsis Information Resource (TAIR, Ohio State University, OH). Homozygotic lines were obtained by selection for three or more generations and analysis of segregation ratios. Lack of gene expression in the mutants was verified by RT-PCR before use.

A MYC-coding sequence was fused in-frame to the 5′ end of the CCA1-coding sequence or to the 3′ end of the LHY-coding sequence, and the gene fusions were subcloned into the pBA002 vector under the control of the Cauliflower Mosaic Virus (CaMV) 35S promoter. The expression constructs were transformed into Col-0 plants, resulting in 35S:MYC-CCA1 and 35S:LHY-MYC, respectively. Overexpression of the transgenes was verified by quantitative real-time RT-PCR (qRT-PCR).

Gene expression assay

Extraction of total RNA samples from appropriate plant materials and qRT-PCR conditions have been described previously [57]. Total RNA samples were pretreated with RNase-free DNase to eliminate contaminating genomic DNA before use.

RNA sample preparation, reverse transcription, and quantitative PCR were conducted according to the criteria that have been proposed to ensure reproducible and accurate measurements [58].

qRT-PCR reactions were performed in 96-well blocks with an Applied Biosystems 7500 Real-Time PCR System (Foster City, CA) using the SYBR Green I master mix in a reaction volume of 20 μl. The PCR primers were designed using the Primer Express Software installed in the system and listed in Additional file 10. The two-step thermal cycling profile employed was 15 s at 94 °C and 1 min at 68 °C. An eIF4A gene (At3g13920) was included in the reactions as internal control to normalize the variations in the amounts of cDNA used. All the qRT-PCR reactions were performed in biological triplicates using RNA samples extracted from three independent plant materials grown under identical conditions. The comparative ΔΔCT method was employed to evaluate relative quantities of each amplified product in the samples. The threshold cycle (CT) was automatically determined for each reaction using the default parameters of the system. The specificity of PCR reactions was determined by melt curve analysis of the amplified products using the standard method installed in the system.

ChIP assay

ChIP assays were performed, essentially as described previously [59], in biological triplicates using three independent plant materials grown under identical conditions. Seven-day-old 35S:MYC-CCA1 and 35S:LHY-MYC transgenic plants grown on MS-agar plates were vacuum-infiltrated with 1 % (v/v) formaldehyde for cross-linking and ground in liquid nitrogen after quenching the cross-linking process. Chromatin preparations were sonicated into 0.5- to 1-kb fragments. An anti-MYC antibody (Millipore, Billerica, MA) was added to the chromatin solution, which was precleared with salmon sperm DNA/protein G agarose beads (Roche, Indianapolis, IN). The precipitates were eluted from the beads. Cross-links were reversed, and residual proteins were removed by incubation with proteinase K. DNA was recovered using the QIAquick PCR purification kit (Qiagen, Valencia, CA). Quantitative PCR was performed to determine the amounts of genomic DNA enriched in the chromatin preparations. The primers used are listed in Additional file 10.

Transient expression assays in Arabidopsis protoplasts

In the reporter vector, a 2-kb promoter sequence of FT gene was transcriptionally fused to the β-glucuronidase (GUS) gene. The GUS reporter construct harboring a mutated CCA1-binding sequence (CBS) within the FT promoter was used to investigate the effects of CBS on the binding of CCA1 and LHY to the FT promoter. The CCA1- and LHY-coding sequences were subcloned under the control of the CaMV 35S promoter in the effector vector. The reporter and effector vectors were cotransfected into Arabidopsis mesophyll protoplasts by the polyethylene glycol (PEG)-calcium transfection method [60]. The CaMV 35S promoter-luciferase construct was also cotransfected as internal control. GUS activity was measured by a fluorometric method as described previously [61]. Luciferase activity assay was performed using the Luciferase Assay System kit (Promega, Madison, WI). GUS activities were normalized by luciferase activities.

Flowering time measurement

Plants were grown in soil at 23 °C under either long days of 24 h or short days of 24 h (8 L: 16D) or 20 h (6.6 L: 13.4D) total duration until flowering. Numbers of rosette leaves at bolting were counted, and 20 countings were averaged for each measurement.

Circadian rhythm measurement

Plants were entrained to long day cycles and then transferred to continuous light conditions for 3 days. To trace the circadian rhythm, whole plant materials were harvested at appropriate zeitgeber time (ZT) points for total RNA extraction. Gene transcript levels were measured by qRT-PCR.

Preparation of recombinant MBP-CCA1 fusion protein

A CCA1-coding sequence was subcloned into the pMAL-c2X Escherichia coli (E. coli) expression vector (NEB, Ipswich, MA) harboring a maltose binding protein (MBP)-coding sequence. Recombinant MBP-CCA1 fusion protein was produced in E. coli Rosetta2 (DE3) pLysS strain (Novagen, Madison, WI). Harvested cells were resuspended in MBP buffer (20 mM Tris-HCl, pH 7.4, 200 mM NaCl, 1 mM EDTA, 10 mM 2-mercaptoethanol, 1 mM PMSF, and protease inhibitor cocktail (Roche, Indianapolis, IN)). Cell lysates were prepared by running three cycles of freezing and thawing followed by centrifugation. The fusion proteins were affinity-purified as described previously [62].

EMSA

EMSA was performed using recombinant MBP-CCA1 fusion protein, as described previously [63]. DNA fragments were end-labeled with γ-32P[dATP] using T4 polynucleotide kinase (Takara, Kyoto, Japan). Labeled probes were incubated with 100 ng of MBP or MBP-CCA1 fusion protein for 30 min at room temperature in binding buffer (10 mM Tris-HCl, pH 7.6, 50 mM NaCl, 1 mM EDTA, 5 mM DTT, 5 % glycerol) supplemented with 100 ng poly(dI-dC) in the presence or absence of competitor DNA fragments. The reaction mixtures were resolved on 6 % non-denaturing polyacrylamide gel at 100 V for 1 h. The gels were dried on Whatman 3 MM paper and exposed to X-ray film.