- © 2012 American Society of Plant Biologists. All rights reserved.
Abstract
The Arabidopsis thaliana early in short days6 (esd6) mutant was isolated in a screen for mutations that accelerate flowering time. Among other developmental alterations, esd6 displays early flowering in both long- and short-day conditions. Fine mapping of the mutation showed that the esd6 phenotype is caused by a lesion in the HIGH EXPRESSION OF OSMOTICALLY RESPONSIVE GENES1 (HOS1) locus, which encodes a RING finger–containing E3 ubiquitin ligase. The esd6/hos1 mutation causes decreased FLOWERING LOCUS C expression and requires CONSTANS (CO) protein for its early flowering phenotype under long days. Moreover, CO and HOS1 physically interact in vitro and in planta, and HOS1 regulates CO abundance, particularly during the daylight period. Accordingly, hos1 causes a shift in the regular long-day pattern of expression of FLOWERING LOCUS T (FT) transcript, starting to rise 4 h after dawn in the mutant. In addition, HOS1 interacts synergistically with CONSTITUTIVE PHOTOMORPHOGENIC1, another regulator of CO protein stability, in the regulation of flowering time. Taken together, these results indicate that HOS1 is involved in the control of CO abundance, ensuring that CO activation of FT occurs only when the light period reaches a certain length and preventing precocious flowering in Arabidopsis.
INTRODUCTION
The integration of complex signals from environmental and endogenous cues enables plants to time the floral transition at the most advantageous moment (Michaels, 2009; de Montaigu et al., 2010; Imaizumi, 2010; Jarillo and Piñeiro, 2011). Plants growing at northern latitudes adapt their developmental program to the varying daylengths and temperatures that occur during the year (Jackson, 2009). Arabidopsis thaliana is a facultative long-day (LD) plant in which flowering time is controlled by a network of six major pathways: information about daylength, low winter temperatures, and growth temperature are mediated by the photoperiod, the vernalization and the ambient temperature pathways, respectively. By contrast, the aging, the autonomous, and the gibberellin pathways act more independently of ambient conditions (Fornara et al., 2010). A potent repressor of flowering, FLOWERING LOCUS C (FLC), integrates signals coming from both the vernalization and the autonomous pathways (Amasino, 2010). Eventually, the whole network converges in the regulation of the floral integrators: FLOWERING LOCUS T (FT), TWIN SISTER OF FT (TSF), and SUPPRESSOR OF OVEREXPRESSION OF CONSTANS1 (SOC1) (Fornara et al., 2010).
The photoperiod pathway comprises several genes, including GIGANTEA (GI), CONSTANS (CO), and FT (Kobayashi and Weigel, 2007; Turck et al., 2008). Mutations in any of these genes cause a delay in flowering mainly under LDs, whereas their overexpression accelerates flowering independently of daylength (Turck et al., 2008). CO is a B-box–type protein that acts in the vascular tissue of the leaves to activate FT and TSF transcription (An et al., 2004; Jackson, 2009; Tiwari et al., 2010). CO may induce FT expression by forming a DNA binding complex with NUCLEAR FACTOR Y/HEME ACTIVATOR PROTEIN proteins (Wenkel et al., 2006; Kumimoto et al., 2010) or ASYMMETRIC LEAVES1 (AS1) (Song et al., 2012) and by binding the FT promoter directly at CO-responsive elements (Tiwari et al., 2010), which are functional cis-elements required for FT expression (Adrian et al., 2010). FT protein, and possibly TSF, are part of the florigen that moves to the shoot apical meristem to induce flowering in response to LDs (Corbesier et al., 2007; Jaeger and Wigge, 2007; Jang et al., 2009).
Plants have developed an intricate molecular mechanism to measure daylength based on the coincidence of an internal rhythm, set by the circadian clock, with an external cue, such as light. The ability to distinguish LDs from short days (SDs) is largely the result of the complex regulation of CO, both at the transcriptional and posttranslational levels, and may have arisen very early in plant evolution (Serrano et al., 2009; Valverde, 2011). Under LDs, CO mRNA shows two peaks of expression, the first following the expression of GI at the end of a LD, when plants are still exposed to light, and the second during the night. Under SDs, only the night peak of CO expression takes place (Suárez-López et al., 2001). The precise timing of CO also requires the degradation of a family of repressors, the CYCLING DOF FACTORs (CDFs), by the F-box proteins FLAVIN BINDING, KELCH REPEAT, F-BOX1 (FKF1), ZEITLUPE (ZTL), and LOV KELCH PROTEIN2 (LKP2), in conjunction with GI (Imaizumi et al., 2005; Sawa et al., 2007; Fornara et al., 2009).
The increased expression of CO in the light under LDs but not SDs is crucial for the promotion of flowering because exposure to light is required for stabilization of CO protein (Valverde et al., 2004; Jang et al., 2008). The high CO transcript levels detected during the dark phase of both LD and SDs do not correlate with CO protein accumulation because the RING finger protein CONSTITUTIVE PHOTOMORPHOGENIC1 (COP1) promotes CO degradation in the dark (Jang et al., 2008; Liu et al., 2008b). Mutations in COP1, encoding a component of an ubiquitin ligase complex, cause extreme early flowering under SDs. This early flowering phenotype is largely dependent on CO activity and correlates with an increase in FT transcription in the cop1 mutant (Jang et al., 2008; Liu et al., 2008b). COP1 and CO interact both in vivo and in vitro, and it has been proposed that COP1 contributes to daylength perception by reducing the abundance of the CO protein during the night (Jang et al., 2008; Liu et al., 2008b; Chen et al., 2010). However, in the morning, CO degradation occurs independently of COP1 (Jang et al., 2008). Therefore, it has been suggested that an unidentified E3 ubiquitin ligase must collaborate in CO degradation during the early part of the day to ensure that CO induction of FT only takes place in LDs (Jang et al., 2008).
In addition to the length of the daily light/dark periods, plants also perceive light quality. Blue and far-red light promote flowering, but red light delays it (Valverde et al., 2004). Far-red light can increase CO protein levels independently of transcription (Kim et al., 2008). Blue light mediates photoperiodic control of the floral initiation by at least three different mechanisms: First, it promotes the interaction of FKF1 and GI necessary for CDFs degradation (Sawa et al., 2007); second, the blue light receptor Cryptochrome 2 (Cry2) prevents GI and CO proteolysis by COP1 (Liu et al., 2008b; Yu et al., 2008; Zuo et al., 2011); and third, Cry2 modulates FT transcription directly (Liu et al., 2008a). Moreover, the red light photoreceptor phytochrome B (phyB) has been implicated in the degradation of CO during the first part of the day (Valverde et al., 2004; Jang et al., 2008).
Genetic screens devoted to the isolation of early flowering mutants have revealed the existence of genes that repress the floral transition (Pouteau et al., 2004). Floral repressors are essential to safeguard against premature flowering, and knowledge of how these repressors interact with the floral promotion pathways is just emerging (Pouteau et al., 2004; Roux et al., 2006; Jarillo and Piñeiro, 2011). Here, we demonstrate that the early in short days6 (esd6) early flowering mutant is caused by a lesion in the HOS1 gene. HOS1 encodes a protein with E3 ubiquitin ligase activity, previously described as a regulator of cold acclimation responses (Lee et al., 2001; Dong et al., 2006a). The early flowering phenotype of hos1 is completely suppressed by co mutations in the Landsberg erecta (Ler) background and notably delayed by co mutations in the Columbia (Col) background. Moreover, we show that HOS1 physically interacts with CO in vitro and in planta and regulates CO protein abundance during the daylight period, indicating the participation of another RING finger–containing protein, in addition to COP1, in the photoperiodic control of flowering time in Arabidopsis. Thus, we propose that HOS1 is required to modulate precisely the timing of CO accumulation and that this regulation is essential to maintain low levels of FT during the first part of the day and, subsequently, a correct photoperiodic response in Arabidopsis.
RESULTS
The esd6 Mutant Is Early Flowering and Displays Pleiotropic Defects in Both Vegetative and Reproductive Development
A recessive mutation that accelerated flowering time, named esd6, was identified by screening a Ler mutagenized population. Plants homozygous for esd6 were selected as early flowering under LD conditions, although the esd6 mutation also accelerated flowering under noninductive SD photoperiods (Figures 1A and 1B, Table 1). Earliness of esd6 was mainly associated with the production of fewer leaves during the adult vegetative phase (Figure 1D) based on leaf trichome distribution (Telfer et al., 1997).
Phenotypic Characterization of the esd6 Mutant.
(A) and (B) Flowering time phenotype of Ler and esd6 plants grown in LD conditions for 23 d (A) or in SD conditions for 54 d (B).
(C) Phenotype of Ler and esd6 plants grown in LD conditions for 35 d.
(D) Histograms comparing the number of juvenile, adult, and cauline leaves in Ler and esd6 plants grown under both LD and SD photoperiods.
(E) Rosette and cauline leaves of Ler and esd6 plants grown in LDs.
(F) Detached Ler and esd6 flowers and siliques from plants grown under LD conditions.
(G) Root elongation in 11-d-old Ler and esd6 seedlings.
(H) Total chlorophyll content (Ct) in Ler and esd6 and phyB-1 mutant seedlings.
Flowering Time of hos1 Double Mutants
Besides the flowering phenotype, esd6 mutant plants also displayed complex pleiotropic alterations of both vegetative and reproductive development. Mutant plants were smaller than the wild type (Figures 1A and 1B) and showed reduced leaf size compared with Ler (Figure 1E). Moreover, esd6 primary root was shorter and produced fewer secondary roots than the wild type (Figure 1G; see Supplemental Figure 1 online). By contrast, stem length was not noticeably affected by the esd6 mutation (Figure 1C). esd6 flowers also displayed some developmental abnormalities, including a reduced flower size in comparison to the wild type (Figure 1F; see Supplemental Figure 1 online). In addition, siliques were ~30% shorter in esd6 mutant than in Ler (Figure 1F; see Supplemental Figure 1 online).
Because the esd6 mutant looked paler than Ler, we decided to measure the total chlorophyll content present in both genotypes (Figure 1H). We included phyB-1 as a control in this experiment, since phyB mutants display reduced chlorophyll accumulation (Reed et al., 1993). As expected, both phyB-1 and esd6 mutants showed less chlorophyll than Ler (Figure 1H), indicating an additional role of ESD6 gene in the regulation of chlorophyll biosynthesis.
The ESD6 Gene Encodes HOS1, an E3 Ubiquitin Ligase
esd6 was identified in a Ler transposon-mutagenized population generated from the cross between two transgenic Ler plants, one containing the Dissociation (Ds) element and the other the transposase gene, capable of mobilizing the Ds element. The esd6 mutant was selected in the F2 population where plants with different phenotypes were observed due to the mobilization of the transposon. The esd6 early flowering phenotype did not cosegregate with the selection resistance gene. For this reason, we considered that the mutation originated from a second mobilization event of the Ds element that left a fingerprint in the genome. Consequently, to understand the molecular function of ESD6, we performed a map-based cloning approach. ESD6 was initially located in the upper arm of chromosome 2, and further linkage analyses allowed us to define a candidate region between C005 and T5I7 molecular markers (see Supplemental Table 1 online), which encompassed eight open reading frames (see Supplemental Figure 2 online). Among these loci, HOS1 (At2g39810) had already been described as a regulator of cold acclimation responses also affecting flowering time (Lee et al., 2001). The sequencing of this transcription unit in esd6 revealed a single nucleotide deletion in the position 2212 (fifth exon), which generated a premature stop codon (Figure 2A). To confirm that esd6 was indeed affecting the same locus as the hos1-1 mutation, we performed an allelism test. The F1 plants derived from the cross between hos1-1, in C24 background, and the esd6 mutant were early flowering, indicating that the two mutations were allelic (Table 1). The esd6 mutant will be referred to hereafter as hos1-2. In addition, we searched for T-DNA insertion alleles within the HOS1 locus and identified the line SALK_069312, which carried an insertion in the fifth exon of the HOS1 gene (Figure 2A). This T-DNA mutant allele was named hos1-3, and an additional allele, hos1-4, was obtained later on in our laboratory during the screening of an ethyl methanesulfonate–mutagenized population of Col plants. The hos1-4 mutation is a deletion of nucleotide 88 from the annotated translational start of HOS1 that generates a premature stop codon in the first exon of the gene (Figure 2A). All hos1 alleles analyzed display an early flowering phenotype both in LD and SD photoperiods, but the fact that they flower earlier under inductive photoperiods indicates that the mutation does not completely abolish the plant photoperiod response (Figure 2B, Table 1).
ESD6 Encodes HOS1.
(A) ESD6/HOS1 gene structure showing the mutations present in the different hos1 alleles. Exons are represented by boxes and introns by a line.
(B) Pictures illustrating the flowering time of hos1 mutants and their respective wild-type genotypes in LD and SD conditions. Plants were grown for 23 d under LD conditions (Top). SD pictures (Bottom) were taken after 64 d for C24 and hos1-1, 58 d for Ler and hos1-2, and 60 d for Col and hos1-3.
The HOS1 gene is around 5.5 kb long, bears nine exons, and encodes a protein of 915 amino acids that contains a noncanonical RING finger domain in the N-terminal region and a putative nuclear localization signal in the C-terminal part (Figure 2A; see Supplemental Figure 3 online). HOS1 is a unique gene in Arabidopsis, and putative orthologs have only been found in plants. The Cys residues present in the RING finger domain are totally conserved between all HOS1 orthologs (see Supplemental Figure 3 online). RING finger domains are found in proteins with E3 ubiquitin ligase activity that participate in the ubiquitin/26S proteasome pathway (Deshaies and Joazeiro, 2009; Vierstra, 2009). Previously, it was shown that Arabidopsis HOS1 can function as an E3 ubiquitin ligase in ubiquitination assays (Dong et al., 2006a).
hos1 Mutations Affect FLC Expression and Have an FLC-Independent Effect in the Control of Flowering Time
The early flowering phenotype of hos1 mutants suggested that HOS1 could be a negative regulator of the floral transition in Arabidopsis. To test this hypothesis, we analyzed the phenotype of double mutants carrying hos1 and different mutations affecting flowering time. It had been previously described that FLC expression levels were reduced in the hos1-1 mutant compared with C24 accession (Lee et al., 2001). In order to check if the hos1 early flowering phenotype was fully dependent on FLC, the effect of the hos1-3 mutation in an flc null genetic background (flc-3) (Michaels and Amasino, 1999) was tested. Both hos1-3 and hos1-3 flc-3 double mutant plants flowered with the same number of leaves, irrespectively of photoperiodic conditions, although the hos1-3 flc-3 plants bolted consistently earlier than hos1-3 under LD (Figure 3A, Table 1). This result may indicate that there is no additional effect of flc null mutation on the acceleration of flowering time caused by hos1. Besides, both hos1-3 and hos1-3 flc-3 plants flowered clearly earlier than flc-3 plants under both LD and SD conditions (Figure 3A, Table 1), indicating that the effect of the hos1 mutation on flowering time could not be exclusively dependent on FLC activity and that there is an FLC-independent effect responsible for the early flowering phenotype of hos1. To find out whether FLC expression was altered in the hos1 mutant alleles isolated in different backgrounds, RT-PCR analyses were performed in hos1-1, hos1-2, and hos1-3 mutants and the corresponding wild-type genotypes. Consistent with previous results (Lee et al., 2001), in all hos1 mutants assayed, FLC transcript was clearly downregulated (Figure 3B); therefore, it cannot be ruled out that this change in FLC expression has an effect on the early flowering time of the hos1 alleles.
hos1 Mutations Downregulate FLC Expression and Have an FLC-Independent Effect in the Control of Flowering Time.
(A) Flowering time phenotype of hos1-3 flc-3 double mutant in LD (Top) and SD (Bottom) conditions.
(B) Analysis of the expression of FLC in 14-d-old hos1 mutant seedlings and their corresponding wild-type genotypes. FLC expression was monitored by RT-PCR using 22 cycles for C24 and hos1-1 and 28 cycles for Ler, hos1-2, Col, and hos1-3. For the UBQ10 control, 22 cycles were used.
(C) Flowering time phenotype of hos1-3 fve-3 double mutant plants grown in LD (Top) and SD (Bottom) conditions.
(D) Flowering time phenotype of hos1-3 plants bearing an active allele of FRI in LDs (Top) and SDs (Bottom).
Dominant alleles of FRIGIDA (FRI) confer a vernalization requirement that delays flowering through the upregulation of FLC (Johanson et al., 2000). To find out the genetic relationship between HOS1 and FRI, the mutant hos1-3 was crossed with a Col plant bearing an active FRI allele introgressed from the San Feliu-2 (Sf-2) accession (Lee and Amasino, 1995). Under LD conditions, the hos1-3 FRI Sf-2 line showed an additive phenotype, the FRI late-flowering phenotype being only partially suppressed by hos1-3 (Figure 3D, top panel, Table 1). This suggests that HOS1 and FRI do not regulate FLC expression through the same pathway in LDs. By contrast, the hos1-3 FRI Sf-2 plant flowered with approximately the same number of leaves as the Col FRI Sf-2 plants under SD conditions, abolishing the effect of the hos1 mutation (Figure 3D, bottom panel, Table 1).
Because HOS1 is involved in cold signal transduction (Lee et al., 2001) and vernalization regulates FLC expression (Amasino, 2010), we hypothesized that HOS1 could be controlling FLC transcript levels through the vernalization pathway. To analyze this, we generated combinations between hos1 and two other mutants impaired in the vernalization response, vernalization1 (vrn1) and vernalization-insensitive3 (vin3) (Levy et al., 2002; Sung and Amasino, 2004), both in late flowering backgrounds that allowed us to observe the acceleration of flowering due to the vernalization treatment. The hos1 mutation did not impair the acceleration of flowering caused by vernalization when combined with the late flowering fca-1 or FRI Sf-2 plants (see Supplemental Table 2 online). Besides, no difference in flowering time was found for the hos1-2 vrn1-2 fca-1 triple mutant grown after either 1 or 4 weeks of vernalization treatment (see Supplemental Table 2 online). The same result was observed for hos1-3 vin3-4 carrying an active FRI allele, as both 1- and 4-week-vernalized plants flowered with approximately the same number of leaves (see Supplemental Table 2 online). Thus, we conclude that HOS1 does not regulate FLC expression through the vernalization pathway.
Considering that the autonomous pathway also converges on the regulation of FLC expression, the flowering phenotype of double mutants combining hos1 and mutations in representative autonomous pathway genes was analyzed, in particular the hos1-3 fve-3, hos1-2 fca-1, and hos1-3 fld-1 double mutants. Under LD, these double mutant plants showed an additive phenotype because the late-flowering phenotype of autonomous pathway mutants was only partially suppressed by hos1 (Figure 3C, top panel, Table 1). By contrast, under SDs, flowering time of these double mutants was very similar to the one displayed by the autonomous pathway mutants, as they produced only a few leaves less than fve-3, fca-1, and fld-1 respectively (Figure 3C, bottom panel, Table 1).
Altogether, these results indicate that the hos1 mutation cannot accelerate flowering in SDs when combined with genetic backgrounds that have very high FLC expression levels, such as mutations of the autonomous pathway or active alleles of FRI. By contrast, under LDs, the repressive effect of HOS1 on flowering time may be mediated by additional pathways that remain inactive in SDs.
The E3 SUMO ligase SIZ1 promotes FLC expression by repressing the autonomous pathway gene FLD (Jin et al., 2008). Besides, SIZ1 stabilizes the ICE1 protein, which has been implicated in the regulation of freezing tolerance in Arabidopsis (Miura et al., 2007). Because it had been described that ICE1 was also targeted by HOS1 (Dong et al., 2006a), we checked the genetic relationship between hos1 and siz1 mutants. Flowering time of siz1 plants relative to the wild type was slightly earlier under LDs and substantially earlier under SDs (Jin et al., 2008) (Table 1). When we combined siz1-2 with the hos1-2 mutation, the double mutant flowering time resembled that of hos1-2 in LDs but was earlier than any of the parental lines in SDs, suggesting a synergistic genetic interaction between both loci (Table 1).
The Early Flowering Phenotype of hos1-2 Requires a Functional CO Gene
We also analyzed the phenotype of double mutants carrying hos1 and mutations in genes representative of the photoperiod pathway, such as CRY2/FHA, GI, and CO, which delay flowering mainly under LDs (Koornneef et al., 1998). While hos1-2 fha-1 and hos1-2 gi-3 double mutants showed an additive flowering phenotype between hos1-2 and fha-1 and gi-3 late flowering mutants, the genetic interaction observed between hos1-2 and co-2 was completely different (Figure 4, Table 1). Under LDs, the hos1-2 mutation did not accelerate flowering time when it was combined with co-2 (Figure 4C, Table 1); indeed, hos1-2 co-2 plants flowered with the same number of leaves as co-2 mutants, indicating that the early flowering phenotype of hos1-2 has a strong requirement for a functional CO gene. However, under SD conditions, hos1-2 co-2 flowered as early as hos1-2 (Table 1), given that co mutations do not delay flowering under this photoperiodic condition. These genetic results suggest that HOS1 is involved in the photoperiodic control of flowering time as a negative regulator of CO under LDs.
Genetic Analyses of hos1 and Mutations in Photoperiod Pathway Genes.
Flowering time phenotype of hos1-2 fha-1 (A), hos1-2 gi-3 (B), and hos1-2 co-2 (C) double mutants grown in LD conditions. Pictures were taken after 22, 43, and 38 d, respectively.
FKF1 is an F-box protein (Imaizumi et al., 2005) that mediates the cyclic degradation of CDF proteins, which are repressors of CO expression (Imaizumi et al., 2005; Fornara et al., 2009). To study if there was any genetic interaction between FKF1 and HOS1, the double mutant hos1-3 fkf1-1was analyzed and it showed an additive phenotype between the late flowering time of fkf1-1 and the early flowering phenotype of hos1-3 in LDs (Table 1). This result indicates that HOS1 does not participate in the FKF1 transcriptional regulatory pathway that controls CO expression.
hos1 Mutants Show an Altered Pattern of FT Expression
FLC represses the expression of the floral integrators FT and SOC1, while the photoperiod pathway activates FT and SOC1 expression through CO (Yoo et al., 2005; Searle et al., 2006; Turck et al., 2008). Because hos1 mutations showed downregulation of FLC expression (Figure 3B) and the co-2 mutation was epistatic to hos1-2 under LDs (Figure 4C), we decided to check the genetic relationship between HOS1 and the floral integrators FT and SOC1. The hos1-2 ft-1 double mutant showed a similar flowering phenotype to ft-1 under LD conditions, suggesting a strong requirement for FT for the hos1 early flowering phenotype (Figure 5A, Table 1). By contrast, the hos1-2 soc1-1 double mutant was additive between the parental lines in both LD and SD conditions (Figure 5A, Table 1). This result is in accordance with the epistasis observed between co-2 and hos1-2, considering that FT is the main target of CO under LDs (Yoo et al., 2005). To check whether the whole effect of HOS1 on flowering time was through FT and SOC1, we generated the triple mutant hos1-2 ft-1 soc1-1. Flowering time analysis showed that the triple mutant was slightly earlier than ft-1soc1-1 (Figure 5A, Table 1). This result indicates that the early flowering phenotype of hos1 mutation requires functional FT and SOC1 genes, although we cannot rule out that HOS1 could regulate other protein(s) involved in the control of flowering time.
The Early Flowering Phenotype of hos1 Depends on FT and SOC1 Functional Proteins and hos1 Mutation Alters the Pattern of Expression of FT.
(A) Flowering time phenotype of hos1-2 soc1-1, hos1-2 ft-1, and hos1-2 ft-1 soc1-1 triple mutant plants grown in LD conditions.
(B) HOS1 expression pattern over a 24-h time course in Ler seedlings grown for 8 d in LDs and for 16 d in SDs. Samples were harvested every 4 h after dawn. HOS1 expression was monitored by RT-PCR using 20 cycles.
(C) and (D) Expression analysis of different flowering time genes over a 24-h time course in Ler and hos1-2 seedlings grown for 8 d in LDs and 16 d in SDs.
(C) Samples were harvested every 4 h after dawn. Q-PCR analysis of CO, FT, and SOC1 expression in LD conditions.
(D) Same as (C) but Ler and hos1-2 seedlings grown in SD conditions. Relative expression levels were normalized to ACT2 expression.
To deepen our understanding of the genetic relationship observed between HOS1 and the photoperiod pathway, we performed a time course expression analysis over a 24-h period in Ler and hos1-2 seedlings both in LDs and SDs. First of all, we demonstrated that HOS1 transcript did not show a diurnal oscillation in Ler background (Figure 5B). Next, we analyzed the temporal expression pattern of CO, FT, and SOC1 genes using RT-PCR (see Supplemental Figure 4 online) and quantitative real-time PCR (Q-PCR) (Figures 5C and 5D) approaches. As previously reported, CO transcript level in Ler background was high at dawn and dusk and during the night, remaining low for the rest of the light period of the day (Suárez-López et al., 2001). In hos1-2, the same pattern of CO expression was observed, indicating that hos1 mutation did not affect significantly the levels or the expression profile of the CO transcript (Figures 5C and 5D). However, the expression pattern of the floral integrator FT was clearly altered in the hos1-2 mutant compared with the wild type both in LD and SD conditions (Figures 5C and 5D). FT transcript usually shows a peak of expression at dusk in LDs (around zeitgeber time (ZT) 16), following the evening increase observed in CO mRNA. In the hos1-2 mutant, a peak of FT expression in the subjective morning, mainly at ZT4, but also at ZT8 was observed, when the CO transcript levels are barely detectable in the mutant (Figure 5C). To check whether this alteration was due to a specific developmental stage of the plant or if it relied on the genetic background, FT transcript levels were analyzed at ZT4 and ZT8 in Col and hos1-3 plants harvested 8, 10, 12, and 15 d after germination. In every single stage tested, FT expression was higher in hos1-3 in relation to Col in the first part of the day (see Supplemental Figure 5 online). In SDs, an increased FT expression in the hos1-2 mutant, starting to rise at ZT8 and peaking at ZT12 was observed, which may explain the early flowering phenotype displayed by the mutant under noninductive photoperiods (Figure 5D). A small but consistent increase in SOC1 expression was also detected in hos1-2 plants grown in LD photoperiods in comparison to the wild type (Figure 5C). Thus, we conclude that CO transcript levels are not modified substantially by the hos1 mutation and that HOS1 is required to repress the expression of FT in the first part of the day in LD. Considering that HOS1 has an E3 ubiquitin ligase activity, we speculated that it may be involved in the degradation of proteins that regulate FT expression. Both the genetic analysis and the expression assays suggested that this protein could be CO.
HOS1 Interacts in Vitro and in Vivo with CO and Regulates Its Abundance
Given the proposed epistatic interaction between co-2 and hos1-2 mutants and considering that CO transcript levels were not affected in the hos1-2 mutant, we decided to analyze whether there was a physical interaction between CO and HOS1. For this purpose, we conducted in vitro pull-down experiments using Maltose Binding Protein (MBP)-HOS1 and in vitro–translated CO protein. As shown in Figure 6A, MBP-HOS1, but not MBP alone, was able to interact with CO protein. Whether the interaction between CO and HOS1 also occurred in vivo was tested using bimolecular fluorescence complementation (BiFC). For BiFC, the N terminus of yellow fluorescent protein (YFP) was cloned upstream of HOS1 (YFN-HOS1), and the C terminus of YFP was fused to CO (YFC-CO). By Agrobacterium tumefaciens coinfiltration, these constructs were transiently expressed in abaxial epidermal cells of tobacco (Nicotiana benthamiana) and Arabidopsis leaves (Voinnet et al., 2003). Reconstitution of YFP fluorescence was examined by confocal microscopy 2 d after transient coexpression of the protein pairs. Yellow fluorescence in the nucleus was detected with coexpression of YFC-CO and YFN-HOS1, but no yellow fluorescence was observed when YFC-CO was coexpressed with YFN-AKINβ or when YFC-AKIN10 was coexpressed with YFN-HOS1, as negative controls (Figure 6B). As a positive control, the interaction between amino and carboxy parts of AKINβ and AKIN10 Sucrose nonfermenting1 (SNF1)-related protein kinases (Ferrando et al., 2001) was tested (Figure 6B). The CO–HOS1 interaction was observed in conspicuous nuclear speckles, which have been often associated with foci of proteasome degradation, as previously described for the interaction between CO and COP1 (Jang et al., 2008). These results demonstrate that CO and HOS1 colocalize and physically interact in the nuclei of plant cells.
HOS1 Interacts with CO and Regulates Its Abundance.
(A) HOS1 and CO interact in vitro. A Coomassie blue-stained SDS-PAGE showing MBP (42 kD) and MBP-HOS1 fusion protein (147 kD) expressed in E. coli BL21 Rosetta strain and purified on amylose resin is shown in the top panel. The bottom panel shows the result of a pull-down assay with MBP and MBP-HOS1 proteins incubated with [35S]Met-labeled CO protein. Retained CO protein was visualized after autoradiography of the dried gel.
(B) HOS1 interacts with CO in planta. BiFC assay coexpressing the C terminus of YFP fused to CO (YFC-CO) and the N terminus of YFP to HOS1 (YFN-HOS1) in tobacco and Arabidopsis leaves. Yellow fluorescence in the nucleus indicated interaction. Negative and positive controls were included in the assay.
(C) In vivo coimmunoprecipitation between HOS1 and CO. GFP-HOS1 protein was immunoprecipitated from 35S:GFP-HOS1 35S:CO (HC) and 35S:GFP-HOS1 (H) plants employing GFP antibody-agarose columns and detected with anti-CO antibodies (elution HC, Top). The arrow marks both the coimmunoprecipitated CO protein and input. Antibodies against ubiquitin recognized, in the same blot, a higher molecular mass band (asterisk) corresponding to an ubiquitinated CO form (elution HC, Bottom). Controls included input (1/50) from HC and H samples as well as column flow-through and washes. Protein markers are shown on the right.
(D) Noninvasive in vivo luciferase imaging of 35S:CO–LUC 6-2 and 35S:CO–LUC 6-2/hos1-3 seedlings. Pictures show 7-d-old seedlings grown in LDs 3 h after the lights are on. At this time, more CO-LUC protein accumulates in hos1 mutant (below right) than in the wild type (below left).
(E) Quantification of the luciferase activity in 35S:CO–LUC 6-2 (blue bars) and 35S:CO–LUC 6-2/hos1-3 (gray bars) seedlings expressed as LUC counts per second (LCPS) in serial dilutions of fresh tissue in Steadylite Plus Reagent (mg/mL). Error bars represent sd.
(F) Immunoblot showing CO protein levels during a 24-h time course in nuclear extracts from Col and hos1-3 plants grown under LDs. Numbers above each lane represent hours after dawn that the sample was harvested. Histone H3 was used as a loading control. Relative quantification of each band compared with the control is expressed below the top panel (α-CO).
The interaction between HOS1 and CO proteins was further explored by in vivo coimmunoprecipitation analysis. For this experiment, we generated 35S:HOS1-GFP (for green fluorescent protein) transgenic plants in Col background, and we also introduced this construct into 35S:CO plants. Nuclear proteins isolated from 2-week-old 35S:HOS1-GFP 35S:CO and 35S:HOS1-GFP plants, grown in LD and harvested at ZT4 in the presence of MG132 proteasome inhibitor, were immunoprecipitated after incubation with an anti-GFP antibody (Figure 6C). The eluted purified proteins were blotted and detected with anti-CO antibody. As shown in Figure 6C, CO antibody recognized a clear signal in nuclear extracts from 35S:HOS1-GFP 35S:CO (Figure 6C, elution HC, marked by an arrow) plants but not in control extracts from the 35S:HOS1-GFP (Figure 6C, elution H) plants. Given their apparent molecular mass and the comigration with the input CO protein (Figure 6C, marked with an arrow), which included 1/50 protein amount of the immunoprecipitated fraction (Figure 6C, input HC), the recognized bands must correspond to CO protein. In this blot, a higher molecular mass CO form (Figure 6C, asterisk, top panel) could also be detected. Because HOS1 has been described as a RING finger E3 ligase (Dong et al., 2006a), the same blot was probed with antiubiquitin antibody, which detected only the upper immunoprecipitated band (Figure 6C, asterisk, bottom panel). This result demonstrates that a fraction of the immunoprecipitated CO protein was ubiquitinated in vivo, further supporting the association between both proteins and pointing to a role of HOS1 in the proteasome-dependent degradation of CO.
It has been reported that HOS1 has autoubiquitination ligase activity in vitro and that it can also mediate the ubiquitination and degradation of ICE1 transcription factor (Dong et al., 2006a). To further analyze whether HOS1 may also regulate CO degradation in vivo, a construct constitutively expressing CO fused to luciferase (LUC) was transformed into wild-type Arabidopsis plants. One representative line, 35S:CO–LUC 6-2, displaying an early flowering phenotype, was crossed into hos1-3. The 35S:CO–LUC 6-2/hos1-3 plants flowered earlier than either the hos1-3 mutant or the 35S:CO–LUC 6-2 plants, indicating that the CO–LUC construct was fully functional (see Supplemental Table 3 online). Using in vivo imaging of LUC fluorescence, we found that under LDs the CO protein levels were substantially lower in the wild type than in the hos1 mutant background three hours after dawn (ZT3) (Figure 6D), suggesting that the degradation of CO that occurs in the wild type is impaired in the hos1 mutant. Quantification of LUC activity corroborated that CO protein accumulated to higher levels in the hos1 mutant than in the wild-type plants (Figure 6E). This accumulation of CO protein observed at ZT3 correlates with the early peak of FT expression present in the hos1 mutant (Figures 5C and 5D; see Supplemental Figure 4 online). To further assess the role of HOS1 in CO regulation, we performed immunoblot assays to detect CO protein in nuclear extracts from wild-type and hos1-3 plants grown under LD photoperiods. In these immunoblots, CO protein was present at lower abundance in the wild type than in the hos1-3 mutant plants, particularly during the daylight period (Figure 6F). From these data, we conclude that HOS1 is involved in the photoperiodic regulation of flowering through the modulation of CO protein levels in vivo.
HOS1 Is a Nuclear-Localized Protein That Acts in the Phloem to Regulate Photoperiodic Flowering
HOS1 protein is ubiquitously expressed in all plant tissues (Lee et al., 2001). Computer analysis of the HOS1 amino acid sequence predicted a nuclear localization signal in the C terminus of the protein (see Supplemental Figure 3 online). Previous reports localized HOS1 into the cytoplasm of transgenic Arabidopsis seedlings overexpressing a HOS1-GFP construct, grown under dark conditions at normal growth temperature. However, in response to low temperature treatments, HOS1-GFP accumulated in the nucleus (Lee et al., 2001). To determine whether the subcellular localization of HOS1 was altered by light/dark conditions, we overexpressed a HOS1-GFP construct in the hos1-3 mutant. The homozygous line 35S:HOS1-GFP/hos1-3 4-1-4 showed a delay in flowering time when compared with hos1-3, indicating that the fusion protein was functional in the repression of flowering (Figure 7E). Subsequently, we grew 35S:HOS1-GFP/hos1-3 4-1-4 transgenic plants at 22°C under both continuous light and dark and analyzed GFP fluorescence in root cells by confocal microscopy. As shown in Figures 7A to 7C, HOS1-GFP was clearly targeted to the nucleus; whereas the fluorescence signal could be detected throughout the whole organelle, a significant fraction was localized to the nuclear envelope, independently of the light growing conditions. These results are in agreement with recent observations implicating HOS1 as an interactor of RNA export factor 1 (RAE1) nucleoporin (Tamura et al., 2010). Moreover, the nuclear localization of HOS1-GFP is also consistent with the results of the BiFC assay described above (Figure 6B) and with the detection of CO and other HOS1 targets in the nucleus (Valverde et al., 2004; Dong et al., 2006a).
HOS1 Is a Nuclear-Localized Protein That Is Required in the Phloem to Regulate Photoperiodic Flowering.
Localization of HOS1-GFP in the root cells of 10-d-old 35S:HOS1-GFP/hos1-3 plants analyzed under confocal microscopy.
(A) Plants grown under continuous light.
(B) Plants grown in darkness.
(C) A representative nuclear image of a light-grown seedling showing 4′,6-diamidino-2-phenylindole staining (DAPI; Top), GFP fluorescence (Middle), and the merge of both images (Bottom).
(D) Tissue-specific expression of HOS1 in phloem companion cells and in the shoot apical meristem of hos1 transgenic plants. Flowering time phenotype of Col, hos1-3, SUC2:HOS1/hos1-3, and KNAT1:HOS1/hos1-3 plants in LDs.
(E) Quantification of flowering time of Col, hos1-3, 35S:HOS1-GFP/hos1-3, SUC2:HOS1/hos1-3, and KNAT1:HOS1/hos1-3 plants grown in LD conditions. Total number of leaves was scored in ~10 plants and is represented as mean ± sd.
Given the described interaction between CO and HOS1 (Figure 6) and because CO acts in the phloem companion cells to activate FT transcription (An et al., 2004), we tested whether HOS1 may also regulate flowering when specifically expressed in phloem tissue. Expression of HOS1 under the control of the phloem companion cell-specific SUCROSE H+ SYMPORTER2 (SUC2) promoter (Imlau et al., 1999) in hos1 mutant plants delayed their flowering time and caused partial complementation of the early-flowering phenotype of hos1 mutant (Figures 7D and 7E). By contrast, expression of HOS1 under the shoot apical meristem specific promoter of the KNOTTED-LIKE FROM ARABIDOPSIS THALIANA1 (KNAT1) gene (Lincoln et al., 1994) in hos1 mutant plants did not alter their flowering time (Figures 7D and 7E), demonstrating that HOS1 expression is required in the phloem companion cells, where CO is expressed, to repress flowering.
HOS1 Interacts Synergistically with COP1 in the Control of Flowering Time
COP1 E3 ubiquitin ligase is involved in the degradation of CO protein during the night (Jang et al., 2008; Liu et al., 2008b). However, CO degradation in the morning occurs independently of COP1 (Jang et al., 2008), and for this reason, we speculate that HOS1 may be involved in this process. In our conditions, cop1-4 mutants flowered dramatically earlier than wild-type and hos1 plants under SDs. However, under LDs, cop1-4 mutants flowered earlier than Col but later than hos1 plants (Figure 8A, Table 1). To test the effect on flowering time of abolishing the activity of both HOS1 and COP1, we combined hos1 and cop1 mutations. Interestingly, the hos1-3 cop1-4 double mutant flowered earlier than both parents in LDs and SDs, displaying the same number of leaves in both photoperiodic conditions (Figure 8A, Table 1). This result indicates that the combination of both mutations renders a plant with a complete loss of photoperiod sensitivity and that HOS1 and COP1 genes are functionally related in the control of flowering time.
Genetic Interaction between HOS1 and COP1 in the Control of Flowering Time.
(A) Flowering time phenotype of hos1-3 cop1-4 double mutant in LD (Top) and SD (Bottom) conditions.
(B) Flowering time phenotype of double and triple mutant combinations between hos1-3 co-10 cop1-4 mutants grown in LD conditions.
(C) Comparison of flowering time phenotype between LD-grown Col and hos1-3 plants bearing a 35S:CO transgene.
To further investigate the genetic interaction between HOS1, COP1, and CO genes in controlling Arabidopsis flowering time, a hos1 cop1 co triple mutant was generated and its flowering time was compared with that of hos1 co and cop1 co double mutants (Figure 8B, Table 1). Under LDs, the hos1-3 co-10 and the cop1-4 co-10 plants flowered with 12 and 10 leaves less than co-10, respectively, indicating that part of the early flowering phenotype of the hos1-3 and cop1-4 mutants in Col background occurs independently of CO. This result appears to be in contrast with the epistatic genetic relationship observed between hos1-2 and co-2 alleles in Ler background (Figure 4C, Table 1) and can be explained because the hos1 mutation causes the downregulation of FLC expression (Figure 3B) and FLC is expressed at higher levels in Col than in Ler (Michaels and Amasino, 1999). Besides, the hos1 cop1 co triple mutant flowered with 15 leaves more than the hos1 cop1 double mutant under LD conditions, demonstrating that the co mutation notably delays the hos1 cop1 early flowering phenotype in Col background. Interestingly, the hos1 cop1 co triple mutant formed six and nine leaves less than hos1 co and cop1 co double mutants, respectively (Figure 8B, Table 1), confirming the existence of a synergistic genetic interaction between hos1 and cop1, even in the absence of CO.
Because HOS1 seems to exert an effect as a negative regulator of CO, we tested whether the extremely early flowering of 35S:CO plants (Simon et al., 1996) could be further accelerated by the hos1-3 mutation. To test this hypothesis, the 35S:CO transgene was introduced into wild-type Col and into hos1-3 mutant plants. Although the number of leaves at flowering for both transgenic plants was very similar, we observed that the hos1-3 35S:CO plants bolted consistently earlier than the Col 35S:CO (Figure 8C, Table 1), supporting a role for HOS1 in repressing the promotion of flowering mediated by CO.
DISCUSSION
In many plants, changes in daylength regulate the transition from vegetative growth to flowering, and plants altered in the daylength-sensing mechanism cannot time flowering properly in natural environments (Wilczek et al., 2009). In this work, both genetic and molecular approaches demonstrate that HOS1 is involved in the photoperiodic control of flowering time. The esd6/hos1 mutant was identified through a screen designed to isolate early flowering mutants in Arabidopsis. The characterization of these mutants contributes to unveiling of the mechanisms of action of genes involved in the repression of the floral transition and suggests that a large number of genes participate in this process (Pouteau et al., 2004; Jarillo and Piñeiro, 2011). In addition to precocious flowering, the hos1 mutant showed pleiotropic alterations of leaf, flower, and root development, similar to those displayed by other early flowering mutants (Martin-Trillo et al., 2006; del Olmo et al., 2010).
In Arabidopsis, the flowering response due to changes in photoperiod relies on the interaction of light with the circadian clock–regulated rhythmic expression of CO (Suárez-López et al., 2001). Besides this transcriptional regulation, light-dependent regulation of CO protein stability has also been described (Valverde et al., 2004; Kim et al., 2008). We have demonstrated that HOS1 is a nuclear-localized protein that acts in the phloem to repress flowering time (Figure 7) and that it is involved in regulating CO protein abundance in vivo (Figure 6), ensuring that CO activation of FT only occurs at the appropriate times of the day under inductive photoperiods in Arabidopsis. HOS1 has been reported to work as an E3 ubiquitin ligase that mediates the degradation of the transcription factor ICE1 (Dong et al., 2006a); here, we demonstrated that CO interacts in vitro and in planta with HOS1 and that CO coimmunoprecipitates with HOS1 in vivo (Figure 6). In addition, hos1 mutation altered the FT expression pattern in LD, showing a peak of expression in the subjective morning (Figure 5C). These observations suggest that HOS1 could mediate CO degradation during the daylight period through a mechanism involving ubiquitination and that the timing of HOS1 activity is crucial to establish a photoperiodic flowering response (Figure 9). Both the genetic analysis between CO and HOS1 genes (Figure 4) and the expression analyses of CO transcript and CO protein (Figures 5 and 6) support this hypothesis.
Model for HOS1 Function in the Photoperiodic Control of Flowering Time.
CO transcription depends primarily on the circadian clock (thick black line). In the evening, the degradation of CDFs by the GI/FKF1 complex allows CO transcript levels to increase, and CO protein accumulates due to a photoreceptor-mediated repression of COP1. At this time, CO can promote FT expression and induce flowering. During the night, COP1 activity causes rapid degradation of CO protein by the ubiquitination/26S proteasome system. In the daylight period, HOS1 is required to degrade CO. Additional data will be necessary to establish the possible involvement of HOS1 in the mechanism of CO degradation mediated by phyB that has been proposed to operate in the morning.
Other E3 ubiquitin ligases have been proposed to be involved in the control of flowering time (Cao et al., 2008; Vega-Sánchez et al., 2008; Park et al., 2010). In particular, DAY NEUTRAL FLOWERING (DNF) and COP1 have been demonstrated to regulate the precise pattern of CO expression at the transcriptional and the posttranscriptional level, respectively (Jang et al., 2008; Liu et al., 2008b; Morris et al., 2010). DNF is an important regulator of the rhythm of CO expression, but it is not acting through the GI/FKF1/CDFs regulatory mechanism (Morris et al., 2010). Increased CO transcript in the dnf mutant around ZT 4-6 results in an earlier induction of FT under SDs (Morris et al., 2010). In addition, CO protein is degraded in the dark by the SUPPRESSOR OF PHYA-105 1 (SPA1)-COP1 complex (Laubinger et al., 2006; Jang et al., 2008; Liu et al., 2008b). Also, it has been demonstrated that the Arabidopsis CULLIN4 E3 RING ligase bound to Damaged DNA Binding protein 1 (DDB1) interacts with SPA1-COP1 complex to regulate flowering time (Chen et al., 2010). We have demonstrated that HOS1 also interacts genetically with COP1 in the photoperiodic control of flowering time (Figure 8A). Interestingly, hos1 cop1 double mutants are completely insensitive to photoperiod, and co mutations notably delay the early flowering phenotype of the hos1 cop1 double mutant (Figure 8B). This can be interpreted as HOS1 and COP1 being functionally related proteins in the control of flowering time, regulating CO abundance during the day and in the night, respectively (Figure 9). This is consistent with the observation that the absence of both E3 ubiquitin ligases renders plants unable to distinguish between LDs and SDs. It has been proposed that a phyB-dependent mechanism occurring early in the day may promote CO degradation as well, but the E3 ubiquitin ligase(s) involved in this process remains to be identified (Valverde et al., 2004; Jang et al., 2008). Our data are consistent with HOS1 playing a crucial role in preventing increased CO protein levels and FT expression during early hours of the day. Further analyses will be required to establish the possible participation of HOS1 in the proposed phyB-dependent mechanism of CO proteolysis.
The ability to respond to photoperiod enables plants to anticipate variations in environmental conditions that can be predicted to occur periodically each year. In northern latitudes, shortening daylength in autumn is coupled with decreasing temperatures, while longer days are typically associated with warm temperatures. In addition to repressing the floral transition, HOS1 was previously described as a negative regulator of cold signal transduction (Lee et al., 2001). This suggests that HOS1 might function as an integrative link for both responses, allowing plants to discriminate the duration of the day by regulating CO abundance and to respond to cold temperatures by regulating CBF (for C-repeat binding factors) expression through ICE1 degradation (Dong et al., 2006a). Several lines of evidence point to the existence of overlapping pathways for controlling cold stress and flowering time responses in Arabidopsis (Yoo et al., 2007; Seo et al., 2009). The characterization of several mutants altered in cold acclimation responses has uncovered a role of the corresponding genes in flowering time control. The hos9 and sensitive to freezing6 (sfr6) mutants show a late flowering phenotype, while the long vegetative phase1 (lov1) mutant is early flowering (Zhu et al., 2004; Yoo et al., 2007; Knight et al., 2008). These three genes regulate the expression of cold-inducible genes independently of CBFs, and both LOV1 and SFR6 control flowering time through the photoperiod pathway. Other mutants, such as low expression of osmotically responsive genes4 (los4) and atnup160, display an early flowering phenotype and altered CBF expression levels (Gong et al., 2005; Dong et al., 2006b). Moreover, co and gi photoperiod pathway mutants show altered tolerance to freezing temperatures (Cao et al., 2005; Yoo et al., 2007), and fve mutant flowers late and shows increased expression of FLC and CBF genes (Kim et al., 2004). It has been proposed that in warm late spring, SOC1 downregulates CBFs expression and promotes flowering, but in cold early spring or fall, induction of FLC expression by the CBFs delays flowering and confers cold resistance to the plant (Seo et al., 2009). Besides regulating CO stability, HOS1 controls FLC expression (Figure 3B) (Lee et al., 2001), which is also repressed by prolonged exposure to cold temperatures (Amasino, 2010). The positive effect of HOS1 on FLC expression appears to be independent of the vernalization pathway (Bond et al., 2011) (see Supplemental Table 2 online) and is still uncharacterized. Taken together, these results suggest that HOS1, among other genes, may participate in the photoperiod and temperature signal crosstalk, integrating information coming from both pathways and facilitating a proper response to changing environmental conditions. Thus, this E3 ubiquitin ligase is proposed to integrate both environmental signals, specifically targeting for degradation key factors involved in the regulation of each response. Further studies will be necessary for an in-depth understanding of how these pathways modulate each other’s activity to optimize plant adaptation.
METHODS
Genetic Stocks and Growth Conditions
Arabidopsis thaliana mutant seed stocks used were in Ler, Col, and C24 genetic backgrounds and were obtained from the ABRC of Ohio State University (Columbus, OH), the Nottingham Arabidopsis Stock Centre (NASC) in the UK, and personal donations. C24 accession and mutant hos1-1 seeds were kindly provided by Jian-Kang Zhu (Lee et al., 2001). The monogenic mutants used in this work were described previously: fca-1, ft-1, co-2, and gi-3 (Koornneef, 1991); fve-3 (Ausín et al., 2004); flc-3 (Michaels and Amasino, 1999); phyB-1 (Reed et al., 1993); fha-1 (Guo et al., 1998); vrn1-2 fca-1 (Levy et al., 2002); vin3-4 FRI Sf-2 (Sung and Amasino, 2004); fld-1 (He et al., 2003); siz1-2 (Miura et al., 2005); fkf1-1 (Nelson et al., 2000); cop1-4 (Deng et al., 1991); co-10 (Laubinger et al., 2006); soc1-1 (Samach et al., 2000); and the Col FRI Sf-2 line was described by Lee and Amasino (1995).
Plants were grown in plastic pots containing a mixture of substrate and vermiculite (3:1) or in MS (Murashige and Skoog) medium supplemented with 1% (w/v) Suc and 0.8% (w/v) agar for in vitro culture. Controlled environmental conditions were provided in growth chambers at 22°C and 70% relative humidity. Plants were illuminated with cool-white fluorescent lights (~120 μmol m–2 s–1). LD conditions consisted of 16 h of light followed by 8 h of darkness; SD conditions consisted of 8 h of light followed by 16 h of darkness.
Phenotypic Analysis
Total leaf number was scored as the number of leaves in the rosette (excluding cotyledons) plus the number of leaves in the inflorescence at the time of opening of the first flower (Koornneef, 1991). Cauline, adult, and juvenile leaves were scored independently. Rosette leaves lacking abaxial trichomes were considered as juvenile leaves (Telfer et al., 1997). Data are shown as mean ± sd.
Root length was measured at different developmental stages in seedlings grown in MS medium supplemented with 1% (w/v) Suc and 1% (w/v) plant agar in Petri dishes placed vertically.
Total chlorophyll content was calculated as described by Moran (1982).
Map-Based Cloning of esd6 Mutation and Molecular Characterization of the hos1 Alleles
A mapping population was generated by crossing the esd6 mutant, in Ler background, and a Col wild-type plant. The analysis of 550 early flowering plants with several polymorphic molecular markers (see Supplemental Table 1 online) located the esd6 mutation to the upper arm of chromosome 2, between markers C005 and T5I7. Mutations hos1-1, in C24 background, hos1-2, in Ler background, and hos1-4, in Col background, generate premature stop codons in the seventh, fifth, and first exon of the HOS1 locus, respectively. The T-DNA insertion mutant hos1-3, isolated in Col background, was obtained from the NASC (SALK_069312).
Genetic Analysis
Double mutants were constructed by crossing the monogenic hos1 mutants with lines carrying the mutations flc-3, fca-1, fve-3, fld-1, siz1-2, fha-1, gi-3, co-2, co-10, cop1-4, fkf1-1, ft-1, soc1-1, vrn1-2, or vin3-4. Double mutants were isolated from selfed F2 progeny using molecular markers. A derived cleaved-amplified polymorphic sequences marker was designed for the hos1-2 mutation (PCR amplification using 5′-TTTTTACATGGCCGGTTCAGATC-3′ and 5′- GCAATGTAATGTGAAACTAGGCGA-3′ primers followed by BglII digestion). For the hos1-3 mutation, we used 5′-GGTTTCTGGACCGCATATTTC-3′, 5′-GGCTTCTGACCAGAGAGTGTT-3′, and the SALK LB1 primer. hos1-3 was also crossed with lines carrying the FRI Sf-2 allele (Lee and Amasino, 1995) and the 35S:CO transgene (Simon et al., 1996).
Expression Analysis
Total RNA was isolated using TRIzol (Invitrogen-Gibco), and RT-PCR was performed in a range of DNA concentrations demonstrated to be quantitative. Amplified DNA was electrophoresed and transferred to nylon membranes (Martin-Trillo et al., 2006). For RT-PCR analysis, the HOS1-specific primers 5′-TTGTCCTCTATTTGCGTTTGT-3′ and 5′-TCAAATTGGGGAAGAAGTTATG-3′ were designed to amplify the N-terminal part of the HOS1 coding region. The FLC, CO, FT, and SOC1 primers used were described elsewhere (Piñeiro et al., 2003; Lázaro et al., 2008). UBIQUITIN10 (UBQ10) was used as a loading control in these experiments. Detection was done using radioactively labeled specific probes. Q-PCR analyses were performed using FastStart Universal SYBR Green Master (Roche) and protocols and primers already described for analyzing the expression of CO, FT, SOC1, and ACTIN2 (ACT) genes (Chiang et al., 2009; Morris et al., 2010).
In Vitro Pull-Down Assays
The pMAL and the pMAL-HOS1 constructs were expressed in Escherichia coli BL21 Rosetta strain and the proteins, MBP or MBP-HOS1, were purified on amylose resin (New England Biolabs). In vitro transcription/translation CO reactions were performed with the TNT Quick Coupled Transcription/Translation System (Promega) in the presence of [35S]Met (Amersham Biosciences). For pull-down assays, 1 μg of MBP or MBP-HOS1 bound to beads was incubated with 15 μL of the TNT reaction in 200 μL of binding buffer containing 50 mM HEPES, pH 7.4, 1 mM EDTA, 150 mM NaCl, 10% (v/v) glycerol, 0.1% (v/v) Tween 20, and 0.5 mM DTT (Dong et al., 2006a). The mixture was incubated at room temperature for 1 h and then washed five times with washing buffer (50 mM Tris, pH 7.5, 150 mM NaCl, and 0.2% [v/v] Nonidet P-40). Samples were boiled in the presence of Laemmli buffer and analyzed by SDS-PAGE followed by autoradiography.
BiFC Studies
HOS1 and CO complete open reading frame were cloned in pYFN43 and pYFC43 vectors to produce HOS1 fused to the N-terminal part of YFP (YFN-HOS1) and CO fused to the C-terminal part of the YFP (YFC-CO). These constructs were introduced into Agrobacterium tumefaciens strain C58C1. Five-week-old tobacco (Nicotiana benthamiana) or 3-week-old Arabidopsis Nossen (No-0) plants were leaf inoculated with YFC-CO and YFN-HOS1, the negative control pairs (YFC-CO coexpressed with YFN-AKINβ and YFC-AKIN10 with YFN-HOS1), or a positive control (amino and carboxy parts of AKINβ and AKIN10 SNF1-related kinases; Ferrando et al., 2001), following protocols previously described (Voinnet et al., 2003). Fluorescent interactions were visualized under a Leica TCS SP2 confocal microscope set at 550 nm. Images were analyzed employing Leica LCSLite software.
Nuclear Protein Extraction and Immunological Experiments
Nuclei were isolated from Arabidopsis seedlings that were grown on MS plates for 2 weeks and then frozen. Plants were ground with mortar and pestle in the presence of liquid nitrogen and 30 mL of nuclei isolation buffer containing 50 mM MES-KOH, pH 8.0, 1 mM EDTA, pH 8.0, 30% (v/v) glycerol, 5% (w/v) Suc, 50 mM KCl, 10 mM MgCl2, 10 mM PMSF, 0.1% (v/v) Triton X-100, and plant protease inhibitor cocktail (Sigma-Aldrich). The slurry was filtered through 100-μm mesh and centrifuged sequentially at 6,000 rpm for 20 min, 5,000 rpm for 10 min, and 4,000 rpm for 10 min in a Beckman Avanti J-26 XP centrifuge at 4°C in a JA-25.50 rotor, with the supernatant discarded at each step and the pellets resuspended in the same nuclei isolation buffer. The final pellet was resuspended in 1.5 mL of the same buffer omitting the detergent and centrifuged at 2,000g in a microfuge at 4°C. For immunoblot experiments, the nuclei-enriched preparation was disrupted in the presence of 6 M guanidine chlorhydrate with circular stirring at 4°C, sonicated in a Brandson sonifier set at 10 W, and centrifuged at 20,000g for 10 min in a microfuge at 4°C. The supernatant was precipitated with 90% (v/v) ethanol, recentrifuged at the same speed for 10 min, and washed three times in 90% (v/v) ethanol. The final pellet was dried and resuspended in Laemmli loading buffer and loaded into 4 to 12% (w/v) acrylamide gels. Immunoblots were performed as described before, using anti-CO antibodies (Valverde et al., 2004) and anti-H3 antibodies (AbCAM) as loading controls. Immunochemiluminescence signals were visualized and quantified using a ChemiDoc system (Bio-Rad).
For immunoprecipitation assays, nuclei-enriched preparations were suspended in high salt buffer (Sigma-Aldrich Cell-Lytic kit) including 1/100 Sigma-Aldrich plant protease inhibition cocktail for 4 h at 4°C and disrupted by sonication as before. Extracts were centrifuged at 20,000g in a microfuge at 4°C and the supernatant dialyzed against 50 mM MES, pH 8.0, 0.5% (v/v) Nonidet P-40, 10% (v/v) glycerol, 0.5 mM EDTA, pH 8, 1 mM PMSF, 1/100 Sigma-Aldrich plant protease inhibition cocktail, 1 mM DTT, and 10 mM β-mercapthoethanol for 4 h, at 4°C. Protein extracts (100 μg) were incubated with 50 μL of anti–GFP-Trap-A slurry (Chromotek) in the presence of 0.5 mM MG132 proteasome inhibitor at 4°C overnight, and proteins were processed following the instructions of the manufacturer. Eluted samples were run on a 10% (w/v) acrylamide/bis-acrylamide SDS-PAGE gel and immunoblots performed employing anti-CO and anti-ubiquitin antibodies as described before (Valverde et al., 2004).
LUC Activity Assays
A 35S:CO-LUC construct was transformed in Col plants and homozygous lines were established. Several independent transgenic plants exhibiting early flowering phenotype were selected, and one representative line was crossed with hos1-3 plants. For noninvasive in vivo LUC imaging, 10-d-old Col and hos1-3 seedlings harboring the 35S:CO-LUC construct were grown in MS plates and sprayed with 100 μM luciferin (Biotium) 3 h after dawn. The imaging system consisted of the C2400-32 Photon Counting I-CCD video camera (Hamamatsu Photonics) mounted in a dark chamber. Image acquisition and processing were performed with Wasabi software provided by the camera manufacturer.
Quantification of LUC activity was assayed on seedlings grown in the same conditions described above with a MicroBeta TriLux luminometer (Perkin-Elmer). Seedlings were ground in liquid nitrogen and resuspended in Steadylite Plus Reagent (Perkin-Elmer). The luciferase activity was measured as a mean of three independent experiments and expressed as luciferase counts per second in serial dilutions of fresh tissue in Steadylite Plus Reagent (mg/mL).
Subcellular Localization of HOS1
hos1-3 mutant plants were transformed with a 35S:HOS1-GFP construct, and the selected transgenic plants were grown in MS medium supplemented with 1% (w/v) Suc and 1% (w/v) plant agar in Petri dishes placed vertically. Ten-day-old transgenic plants grown under continuous light or dark conditions were analyzed by confocal microscopy (Zeiss LSM 710). The 4′,6-diamidino-2-phenylindole staining of the nuclei was done at a final concentration of 10 μg/mL with 0.1% (v/v) Tween 20.
Accession Numbers
Sequence data from this article can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: ESD6/HOS1 (At2g39810), CO (At5g15840), SOC1 (At2g45660), FT (At1g65480), FLC (At5g10140), TSF (At4g20370), GI (At1g22770), AS1 (At2g37630), FKF1 (At1g68050), ZTL (At5g57360), LKP2 (At2g18915), COP1 (At2g32950), CRY2 (At1g04400), PHYB (At2g18790), FCA (At4g16280), FVE (At2g19520), VRN1 (At3g18990), VIN3 (At5g57380), FRI (At4g00650), FLD (At3g10390), SIZ1 (At5g60410), UBQ10 (At4g05320), ACT2 (At3g18780), ICE1 (At3g26744), AKINβ2 (At4g16360), AKIN10 (At3g01090), RAE1 (At1g80670), DNF (At3g19140), SPA1 (At2g46340), CUL4 (At5g46210), DDB1A (At4g05420), SFR6 (At4g04920), LOV1 (At2g02450), LOS4 (At3g53110), and At NUP160 (At1g33410).
Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure 1. Flower, Silique, and Root Length Measurement in Ler and esd6.
Supplemental Figure 2. Map-Based Cloning of ESD6.
Supplemental Figure 3. Sequence Comparison of Arabidopsis HOS1 (At HOS1) with the HOS1 Orthologs.
Supplemental Figure 4. Expression Analysis of Different Flowering Time Genes over a 24-h Time Course in Ler and hos1-2.
Supplemental Figure 5. FT Expression Analysis in LD-Grown Col and hos1-3.
Supplemental Table 1. Polymorphic Molecular Markers Used for esd6 Mapping.
Supplemental Table 2. Flowering Time of hos1 and Mutations in Vernalization Pathway Genes.
Supplemental Table 3. Flowering Time of Transgenic Plants Bearing a CO-LUC Construct.
Acknowledgments
We thank Israel Ausín (Centro Nacional de Biotecnología, Consejo Superior de Investigaciones Cientificas, Madrid, Spain) for isolating the esd6 mutant, Rafael Catalá (Centro de Investigaciones Biológicas, Consejo Superior de Investigaciones Cientificas, Madrid, Spain) for providing a Gateway-compatible HOS1 full-length cDNA, Julio Salinas (Centro de Investigaciones Biológicas, Consejo Superior de Investigaciones Cientificas) for granting the access to the luciferase imaging system, Pablo González-Melendi (Centro de Biotecnología y Genómica de Plantas, Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria-Universidad Politécnica de Madrid) for technical support with confocal microscopy, and Juan C. del Pozo and members of his lab (Centro de Biotecnología y Genómica de Plantas, Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria-Universidad Politécnica de Madrid) for helpful discussions. pYFN43 and pYFC43 vectors, and SNF1 protein kinase constructs, used as positive controls in the BiFC assays, were kindly provided by Alejandro Ferrando (Instituto de Biología Molecular y Celular de Plantas, Valencia, Spain). C24 and hos1-1 mutant seeds, as well as a HOS1 expression construct in the pMAL vector, were kindly provided by Jian-Kang Zhu (University of California, Riverside, CA). The 35S:CO construct, the plasmids containing the SUC2 and KNAT1 promoters, and co-10 cop1-4 seeds were kindly donated by George Coupland (Max Planck Institute, Cologne, Germany), and the 35S:CO-LUC plasmid was kindly provided by Hong-Quan Yang (Shanghai Institute for Biological Sciences, Chinese Academy of Science, Shanghai, China). This work was supported by Grant BIO2008-00351 to J.A.J., Grant BIO2007-61215 to M.P., Grants CSD2007-00057 and BIO2010-15589 to J.A.J. and M.P., and Grants BIO2007-61837 and BIO2010-16027 to F.V. from the Spanish Ministerio de Ciencia e Innovación.
AUTHOR CONTRIBUTIONS
M.P. and J.A.J. designed the research. A.L. performed all the experiments except for the immunoblot assays, the BiFC studies, and the coimmunoprecipitation experiments, which were performed by F.V. All the authors analyzed data and wrote the article.
Footnotes
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 (www.plantcell.org) is: Jose A. Jarillo (jarillo{at}inia.es).
↵[W] Online version contains Web-only data.
- Received December 23, 2010.
- Revised February 1, 2012.
- Accepted February 16, 2012.
- Published March 9, 2012.
References
