The Promoter of the Cereal VERNALIZATION1 Gene Is Sufficient for Transcriptional Induction by Prolonged Cold

The VERNALIZATION1 (VRN1) gene of temperate cereals is transcriptionally activated by prolonged cold during winter (vernalization) to promote flowering. To investigate the mechanisms controlling induction of VRN1 by prolonged cold, different regions of the VRN1 gene were fused to the GREEN FLUORESCENT PROTEIN (GFP) reporter and expression of the resulting gene constructs was assayed in transgenic barley (Hordeum vulgare). A 2 kb segment of the promoter of VRN1 was sufficient for GFP expression in the leaves and shoot apex of transgenic barley plants. Fluorescence increased at the shoot apex prior to inflorescence initiation and was subsequently maintained in the developing inflorescence. The promoter was also sufficient for low-temperature induction of GFP expression. A naturally occurring insertion in the proximal promoter, which is associated with elevated VRN1 expression and early flowering in some spring wheats, did not abolish induction of VRN1 transcription by prolonged cold, however. A translational fusion of the promoter and transcribed regions of VRN1 to GFP, VRN1::GFP, was localised to nuclei of cells at the shoot apex of transgenic barley plants. The distribution of VRN1::GFP at the shoot apex was similar to the expression pattern of the VRN1 promoter-GFP reporter gene. Fluorescence from the VRN1::GFP fusion protein increased in the developing leaves after prolonged cold treatment. These observations suggest that the promoter of VRN1 is targeted by mechanisms that trigger vernalization-induced flowering in economically important temperate cereal crops.


Introduction
Plants growing in temperate regions time flowering to coincide with favourable seasonal conditions. Winter frost can damage cold sensitive reproductive organs, while heat and water stress during summer can reduce fertility, so in temperate regions many plants flower in spring when conditions are optimal. One cue that promotes spring flowering is prolonged exposure to winter coldvernalization [1]. Vernalization-induced flowering is a feature of many plants, including Arabidopsis and economically important temperate cereal crops, such as wheat (Triticum spp.) and barley (Hordeum vulgare), although different genes mediate this seasonal flowering response in these distantly related angiosperms (see [2]).
Alleles of VRN1 that are expressed without cold treatment allow flowering without vernalization [6][7][8]. These ''active alleles'' have been used to breed wheats and barleys that flower without vernalization (spring types), which are grown where vernalization does not occur. Some active alleles of the wheat VRN1 gene have mutations near the transcriptional start site, which might disrupt promoter sequences required to repress transcription prior to winter [8], [22], [23]. Additionally, the first intron of VRN1 contains a broad region that is required to maintain repression of VRN1 prior to winter [10], [24][25][26][27]. Alleles lacking large sections of the first intron are actively expressed and are associated with early flowering without vernalization, whereas some alleles lack smaller segments of the first intron and are associated with moderate increases in VRN1 activity and weaker promotion of flowering [26], [27]. An insertion of a mobile genetic element at the 59 end of the first intron is also associated with active expression of VRN1 without vernalization [28], so the repressive action of the first intron is not related simply to size.
The state of chromatin at the VRN1 locus appears to be an important determinant of activity [29]. Without vernalization, the chromatin at VRN1 has high levels of the repressive histone modification histone 3 lysine 27 tri-methylation (H3K27Me3), which is typically associated with an inactive chromatin state [29]. This modification is found within the first intron of the VRN1 gene and at the start point of transcription, sites that are critical for repression [29]. The presence of H3K27Me3 at these sites might contribute to repression of VRN1 prior to winter [29].
The mechanisms that activate expression of VRN1 in response to prolonged cold are unclear. Although the floral repressor VRN2 is required to delay flowering prior to vernalization [18], cold induction of VRN1 takes place in conditions where VRN2 is not actively expressed and can occur in the absence of the VRN2 gene [11], [16], [19]. So VRN2 seems unlikely to mediate cold-induction of VRN1. The hypothesis that the VEGETATIVE TO REPRO-DUCTIVE TRANSITION 2 gene is repressed by cold to allow increased expression of VRN1 [30] also seems unlikely, since VRT2 expression increases at low temperatures [31]. Furthermore, although VRT2 binds to a sequence motif at the promoter of VRN1 in vitro, this motif is not critical for low-temperature induction [23]. Finally, alleles of VRN1 that lack most of the first intron, which are expressed at high basal levels, are induced by cold [27]. This suggests that cold activation might occur independently of the repressive mechanism that acts at the first intron [27].
Studies in Arabidopsis, and other plants, have demonstrated that activation of transcription by low temperatures can be mediated by cis-acting elements located in the promoters of lowtemperature responsive genes [32]. In this study, we examine whether the promoter of the VRN1 gene can mediate a transcriptional response to prolonged cold, using transgenic barley as a model system.

Results
A VRN1 promoter reporter gene fusion is expressed in the leaves and shoot apex of transgenic barley plants A 2 kb fragment from the promoter of the barley (Hordeum vulgare) VERNALIZATION1 gene (referred to hereafter as VRN1) was fused to the GFP reporter gene ( Figure 1A) and the resulting construct (P VRN1 :GFP) was transformed into barley (cv. Golden Promise). GFP activity was detected in the shoot apex and developing leaves of transgenic barley plants that carry the P VRN1 :GFP construct ( Figure 2). GFP fluorescence increased at the shoot apex at the late vegetative stage ( Figure 2B) and remained active throughout the shoot apex during reproductive development, although expression was lower in the developing florets than in other parts of the inflorescence ( Figure 2D). GFP fluorescence was also observed in leaves ( Figure 2E and 2F).
The promoter of VRN1 is sufficient to mediate transcriptional induction of the GFP reporter gene by prolonged cold The levels of GFP transcript generated from the P VRN1 :GFP construct were assayed by quantitative reverse transcriptase PCR (qRT-PCR) in transgenic seedlings exposed to prolonged cold treatment (28 days at 4u). Expression was assayed in three independent transgenic lines. Transcript levels were higher in seedlings at the end of prolonged cold treatment than in control seedlings grown to an identical stage of development at normal temperatures (4 days at 20u) ( Figure 3A). The degree of induction by cold varied between 1.5 to 4 fold, depending on the transgenic line assayed.
A second reporter gene fusion was constructed by fusing the same 2 kb region of the VRN1 promoter to the GFP reporter gene, with the addition of the first and second exons from the VRN1 gene, separated by a small segment (0.35 kb) of the first intron (P VRN1 EX1-2::GFP, Figure 1B). The spliced version of this translational fusion construct was expressed in transgenic barley plants, although no GFP fluorescence was detected, possibly due to poor translation or instability of the fusion protein produced. Levels of the spliced P VRN1 EX1-2::GFP transcript increased in response to prolonged cold treatment ( Figure 3B), similar to the P VRN1 :GFP construct.

Cellular and tissue localisation of a VRN1::GFP translational fusion in transgenic barley
The VRN1 gene, the promoter and transcribed regions minus the majority of the large first intron (1.1 kb of the 10.7 kb intron was present in the construct), was translationally fused to the GFP reporter gene followed by the 39 UTR of the VRN1 gene. The resulting VRN1::GFP construct ( Figure 4A) was transformed into barley. GFP fluorescence signal was observed in the nuclei of cells ( Figure 4). The nuclear VRN1::GFP fluorescence increased during vegetative shoot apex development and was higher in the predouble ridge and double ridge apices than in the early vegetative shoot apex (Figure 4). At later stages of development, the VRN1::GFP signal was detected throughout the shoot apex, although expression was lower in the developing florets than in other parts of the inflorescence, similar to the P VRN1 :GFP construct ( Figure 4H). VRN1::GFP transcript levels increased in seedlings during prolonged cold treatment, although induction was weaker than that observed in the P VRN1 :GFP lines ( Figure S1). Whereas GFP signal was weak in the developing leaves of seedlings grown without cold treatment, strong fluorescence was observed in the nuclei in cells from the developing leaves of seedlings that had experienced prolonged cold ( Figure 5).
Cold response motifs and a small upstream open reading frame are found in the promoter and 59 untranslated region of the VERNALIZATION1 gene The promoter of VRN1 was scanned for potential cis-acting elements, by searching for sequence motifs similar to known transcription factor binding sites. This identified putative binding sites for C-REPEAT BINDING FACTOR (CBF) and INDUCER OF CBF1 (ICE1) transcription factors within the 2 kb of sequence upstream of the predicted translational start site at the VRN1 gene ( Figure 6A, Figure S2). These transcription factors have been shown to activate the expression of cold-responsive genes in Arabidopsis, and other plants [32]. Binding sites for B-ZIP and ethylene responsive transcription factors were also identified.
Another potential regulatory element identified at the VRN1 locus is a small upstream open reading frame (uORF) encoding 13 amino acids (MAPPPRLRPRLSE) in the predicted 59 leader sequence of the VRN1 mRNA ( Figure 6B). As predicted, the uORF is found in the 59 end of VRN1 cDNA sequences in GENBANK, and this was further confirmed by 59 RACE (Rapid Amplification of cDNA ends). The transcriptional start site, which did not vary between vernalized versus non-vernalized plants is annotated in Figure S2. The presence of a proline rich small open reading frame is a conserved feature of the VRN1 genes of barley and wheat. uORFs are also a feature of VRN1-like genes from other cereals and grasses, upstream of the main open reading frame that encodes the AP1/FRUITFULL-like MADS box transcription factor ( Figure 6B). Small uORFs can influence translation from downstream ORFs by decreasing translational initiation rates or by stalling ribosomes and causing non-sense mediated message decay (see [33]).

An insertion in the VRN box does not abolish induction of VRN1 by prolonged cold
Mutations in a proximal promoter region defined as the VRN box are associated with elevated VRN1 expression and early flowering in wheat (Pidal et al. 2009). To examine whether mutations in the VRN box influence the capacity for lowtemperature induction of VRN1, transcriptional induction of VRN1 by prolonged low-temperature treatment was examined in nearisogenic wheat lines with different alleles of the VRN1 gene: a wildtype allele on the A genome (normal VRN box, full length first intron, parental line) versus alleles with either a 7,222 bp deletion within the first intron (line W15) or a 231 bp insertion in the VRN box on the A genome (line W11). Basal expression from the VRN1 gene on the A genome (VRN-A1) was highest in the W11 line carrying the promoter insertion allele (Figure 7), and was also elevated in the W15 line carrying the intron deletion allele (Students T-tests: P,0.05, P,0.01 versus the parental line with the wildtype allele, for the promoter insertion or intron deletion alleles respectively). Low-temperature induction of the A genome copy of VRN1 occurred irrespective of VRN1 genotype (Figure 7). Thus, mutations that are associated with increased basal expression of the VRN1 gene copy on the A genome (homeoallele) do not prevent low-temperature induction.

Discussion
The 2 kb of DNA sequence upstream of the primary translational start site (MADS box open reading frame) of the VRN1 gene is sufficient to drive expression of the GFP reporter gene in the shoot apex and leaves; organs where VRN1 expression has been detected using quantitative RT-PCR [8], [11], [17]. The distribution of P VRN1 :GFP expression observed in shoot apices and leaves of barley ( Figure 2) is similar to the expression patterns of the VRN1 genes of wheat (T. aestivum, T. monococcum) and oats (Avena sativa) detected by in situ hybridisation [17], [30]. Similarly, at later stages of barley inflorescence development ( Figure 2D), the expression pattern of the P VRN1 :GFP construct is consistent with the results of in situ hybridisation analysis of VRN1 expression, which show expression of VRN1 in meristems and organ primordia of the developing barley spike [34].
Expression of the P VRN1 :GFP and VRN1::GFP reporter genes increased at the shoot apex shortly before the transition to reproductive development (inflorescence initiation). This is consistent with the observation that VRN1 transcript levels increase at the shoot apex of this barley cultivar prior to inflorescence initiation, when assayed by quantitative RT-PCR [31]. Similarly, when the promoter of the maize (Zea mays) orthologue of VRN1 (ZmMADS4) was fused to the b-GLUCURONIDASE reporter gene an increase in reporter activity occurs at the equivalent stage of development in transgenic maize plants [35]. This expression pattern was confirmed by in situ hybridisation [35]. Thus, the expression patterns of the VRN1 promoter reporter genes used in this study are similar to those of the endogenous VRN1 genes from a range of cereals.
The increase in VRN1 expression prior to inflorescence initiation and the distribution of VRN1 expression in the developing inflorescence are consistent with a role for VRN1 in promoting inflorescence meristem or organ identity. The lower VRN1 activity seen in the developing florets ( Figure 2D) suggests that less VRN1 activity is required for the differentiation of floral organs. Other VRN1-like genes, Barley MADS3 (BM3) and BM8, might contribute to the development of these organs, compensating for lower VRN1 activity [34].
In addition to driving basal expression of a reporter gene in the shoot apex and developing leaves, a 2 kb segment from the promoter of VRN1 was sufficient for induction of reporter gene expression by prolonged cold. Based on these observations, and those of previous studies [26][27][28][29], we suggest that the vernalization response of cereals depends largely upon two antagonistic regulatory mechanisms that control transcriptional activity of VRN1: activation of transcription by the promoter, which has a low basal activity but can be activated by cold, counteracted by constitutive repression through the first intron of the gene (Figure 8).
A key question is what mechanisms act at the promoter to mediate cold induction? The barley cultivar used in this study lacks VRN2, ruling out involvement of this gene. Similarly, the first intron is not present in the P VRN1 :GFP reporter gene fusion, which is low-temperature responsive ( Figure 3A), consistent with previous suggestions that the intron is not critical for cold induction [27], [31]. The VRN1 promoter contains the putative VRN box, a region mutated in some active VRN1 alleles of wheat [23]. An Figure 3. Expression of the P VRN1 :GFP and P VRN1 EX1-2::GFP constructs in control versus cold treated seedlings. (A) GFP transcript levels in transgenic seedlings carrying the P VRN1 :GFP construct. Expression was assayed by quantitative reverse transcriptase PCR in control seedlings, germinated in darkness at normal glasshouse temperatures (20 degrees for 5 days), compared to seedlings germinated and grown in darkness to an identical stage of development at low temperatures (4 degrees for 28 days). Expression levels were assayed in three independent transgenic lines. (B) Expression of the P VRN1 EX1-2::GFP, assayed as outlined above, in three independent transgenic lines. Expression is shown relative to ACTIN. Error bars show standard error for a minimum of 3 biological replicates. * indicates P,0.05. doi:10.1371/journal.pone.0029456.g003 insertion in this promoter element (the Vrn-A1a allele described by Fu et al. [24]) did not disrupt low-temperature induction of VRN1 (Figure 7), however, suggesting that while this region can influence basal expression of VRN1 it is not critical for low-temperature induction. Presumably mutations in this region affect a transcriptional repression mechanism that acts separately to any mechanisms that mediate low-temperature induction.
The promoter of VRN1 contains sequence motifs that might be targeted by low-temperature responsive transcription factors, including putative CBF and myc recognition sites ( Figure 6). Binding of low-temperature responsive transcription factors, such as CBFs or the myc INDUCER OF CBF EXPRESSION1 (ICE1), to these putative binding sites might play a role in inducing expression of VRN1 during vernalization. The promoter of the barley VRN1 gene also contains binding sites for B-ZIP transcription factors that might mediate long-day activation of VRN1 in vernalized plants or in some early flowering genotypes [36]. The potential role of these putative regulatory elements will be examined in future studies.
Another potential regulatory mechanism identified in the 59 region of VRN1 is a small upstream open reading frame (uORF). uORFs are widespread in mammalian genes and typically reduce protein expression from larger downstream open reading frames by 30-80% [37]. In humans, the presence of a uORFs in a gene can vary between individuals, causing variation in protein expression from downstream open reading frames and this is the basis for some diseases [37]. The uORF in the 59 of the VRN1 gene is conserved amongst VRN1 orthologues in temperate cereals and related grasses ( Figure 6B). The uORF was not required for the transcriptional response to prolonged cold, since it was not included in the low-temperature responsive P VRN1 :GFP construct, but could potentially influence post-transcriptional regulation of the endogenous VRN1 gene. In some early flowering accessions of the wild wheat T. timopheevi the uORF is deleted ( Figure S3, [38]), although it is unclear whether this contributes to natural variation in flowering behaviour of this wild wheat.
In summary, we have shown that the promoter of VRN1 is able to drive basal expression of the GFP reporter gene in the leaves and shoot apex of transgenic barley plants and is sufficient for induction of reporter gene expression by prolonged cold. Understanding the mechanisms that act at the promoter of VRN1 to mediate the transcriptional response to prolonged cold of this central regulator of flowering can provide further insights into how vernalization induces spring flowering of economically important cereal crops.

Construction of VRN1-reporter gene fusion constructs
The P VRN1 :GFP construct was generated by sub-cloning a 2 kb SpeI-NcoI VRN1 promoter fragment from the Morex BAC clone 631P8 and ligating this to a GFP-nopaline synthase terminator cassette. The resulting promoter-reporter gene cassette was inserted into the pVEC8 T-DNA vector [39]. The P VRN1 EX1-2::GFP construct was generated by first amplifying the promoter, first exon and a small segment of the first intron (59- CACACCGCTCTACTAGTTC-39 and 59-GGCGCGCCCTG-CAGGGAGCGCCAGCTCCGCC-39) of VRN1 from a winter barley (cv. Sonja). This was ligated to a second fragment from inside the first intron into the second exon (59-GGCGCG-CCCGATTATCAAAATGTACCA-39 and 59-CCCGGGTCA-CTTGAAACGAGAACC-39), to generate a construct with 2 kb of promoter sequence, with the first and second (partial) exons separated by a 0.35 kb segment of the first intron. The resulting fragment was ligated to the GFP-nopaline synthase terminator, and inserted into pVEC8. The VRN1::GFP fusion was generated by fusing the 39 UTR of VRN1, amplified from the Morex BAC 631P8 (59-TTTCTAGACAGCCCATGTAAGCGTACTATT-CAG-39 and 59-TTACTAGTAATGGCAGGTGTTCTGTT-TGTTTATG-39), to GFP. The GFP-UTR fragment was then ligated to a fragment amplified from the Morex BAC 631P8, from within the first intron to the last exon, minus the stop codon (59-TTTCTAGACAGCCCATGTAAGCGTACTATTCAG-39 and  59-TTACTAGTAATGGCAGGTGTTCTGTTTGTTTATG-39), with the open reading frames of VRN1 and GFP in frame. Primers included compatible restriction sites to facilitate cloning. The resulting VRN1-GFP-UTR fragment was then ligated to the SpeI-NotI promoter fragment and the resulting reporter gene fusion was inserted into the pVEC8 T-DNA vector. The resulting VRN1::GFP fusion lacks most of the first intron (bp 2557 to 12214 relative to the Strider VRN1 sequence) similar to the HvVRN1-3 allele (bp 2968 to 11687).

Gene expression analysis
Seeds were germinated in soil and grown in darkness in foil covered pots at 20 degrees or 4 degrees. RNA was extracted from seedlings with Spectrum Plant Total RNA Kit (Sigma Aldrich, www.sigma-aldrich.com) following manufacturer's instructions. RNA was treated with 10 units of RQ1 RNase-free DNase (Promega, Madison, WI) for 15 min at 37uC. Total RNA (5 mg) was reverse-transcribed with Super Script III reverse transcriptase (Invitrogen, www.invitrogen.com), according to manufacturer instructions. qRT-PCR was performed in a Rotorgene Q realtime PCR cycler (Qiagen; www.qiagen.com) as described previously [20].
The following primer sets were used for quantitative RT-PCR: HvVRN1: 59-GGAAACTGAAGGCGAAGGTTGA and 59-  The expression of each gene was normalized to ACTIN using the Rotorgene software package (Qiagen, www.qiagen.com), which takes amplification efficiency of each primer set into account for quantification calculations. ACTIN primers have been described previously [19]. All data presented are the average of mRNA levels from three biological repeats, unless stated otherwise, with error bars representing the standard error of the mean.

Rapid Amplification of cDNA ends (RACE)
59 RACE was carried out with the FirstChoice RLM-RACE kit from Ambion (Ambion; www.invitrogen.com/ambion) following manufacturer9s instructions with total RNA from Golden Promise seedlings grown under control conditions or Golden Promise seedlings grown to an identical stage of development at 4 degrees (28 days). Nested PCR was performed with the 59RACE Outer Primer provided by the kit and the gene specific VRN1-R Outer primer (59-GCATTCGTGCATAAGTTGGTTC-39) and for the Inner 59 RLM-RACE PCR we used the 59RACE Inner Primer and the gene specific VRN1-R Inner primer (59-TATTGTCT-CAACCTTCGCCTTC-39). PCR products were sequenced after ethanol precipitation.

Confocal microscopy
Barley apices and leaf primordia were dissected under a binocular dissecting microscope and mounted in water. The plant material was then observed in Leica SP2 confocal laser scanning microscope (Leica Microsystems, Sydney, Australia) equipped with a standard Ar 488 laser excitation. Excitation was at 488 nm, and emission was collected between 500 and 550 nm. Chlorophyll autofluorescence was collected between 650 and 720 nm. Nuclear localisation was confirmed by co-localization with 49,6-diamidino-2-phenylindole (DAPI) stain ( Figure S4). Spikes were fixed in 4% paraformaldehyde in 50 mM phosphate buffer, pH 7.2, overnight, then rinsed in buffer and stained with 5 ug/ml DAPI for 30 sec. After mounting in fresh buffer, GFP was detected with 488 nm excitation, and emission from 505-550 nm collected. DAPI was then detected with 405 nm excitation, and emission from 415-485 nm collected using a Leica SP2 confocal microscope in sequential scanning mode to avoid bleedthrough of DAPI fluorescence into the GFP emission channel.

Production of near-isogenic wheat lines
Wheats carrying different VRN1 alleles were crossed to the Australian wheat cultivar Sunstate (PPD-D1 insensitive allele, VRN-A1 wildtype, VRN1-B1 intron deletion and VRN-D1 intron deletion) and then backcrossed a further 4 generations before homozygous plants were selected for contrasting VRN1 genotypes. The donors of different VRN1 alleles were: AUS7374 for Langdon VRN1 A genome allele with an intron deletion, AUS 2380 ''Extra Transcript levels of the A genome copy of VRN1 were assayed in hexaploid bread wheat (Triticum aestivum) near-isogenic lines that vary for VRN1 genotype. Expression was assayed by quantitative reverse transcriptase PCR, using A genome specific primers, in seedlings germinated in darkness at 20u for 5 days (control) or 4u for 6 weeks (cold). The wheat lines used carry different combinations of wildtype VRN1 alleles on each of the three genomes (a, b or d) or alleles with high basal activity (A1a, A-lang., B or D). The A1a allele (line W11) has a promoter insertion in the VRN box of the VRN1 gene on the A genome, but a full-length first intron [22]. The A-lang. allele (line W15), first identified in the Langdon cultivar of the tetraploid wheat Triticum durum [24], has a 7.2 kb deletion within the first intron, but has the wildtype promoter sequence. Similarly, the B and D alleles have deletions within the first intron of the VRN1 genes on the B and D genomes respectively [24]. Expression is shown relative to ACTIN, error bars show standard error for four biological replicates, *** indicates P,0.001. Figure 8. A model for transcriptional regulation of VRN1. Prior to winter, chromatin at the VRN1 locus is maintained in an inactive state by histone modifications deposited by a plant Polycomb Repressor Complex 2 (PRC2), resulting in low transcript levels. When plants are exposed to prolonged cold (snowflake) the promoter of VRN1 becomes more active, leading to increased transcription and higher steady state transcript levels. This triggers a change in the state of chromatin at the VRN1 locus, with a shift towards an active state. doi:10.1371/journal.pone.0029456.g008 Early Blackhull'' for recessive B and D genome alleles or AUS1499 for the VRN-A1a promoter insertion allele. Figure S1 Low-temperature induction of the VRN1::GFP construct.Transcript levels for the VRN1::GFP fusion in transgenic seedlings. Expression was assayed in control seedlings, germinated at normal glasshouse temperatures (20 degrees for 4 days), and compared to seedlings germinated and grown to an identical stage of development at low temperatures (4 degrees for 28 days). Expression levels were assayed in two independent transgenic lines. Error bars show standard error. * indicates P,0.05. (TIF)