Low temperatures induce rapid changes in chromatin state and transcript levels of the cereal VERNALIZATION1 gene.

Transcriptional activation of the VERNALIZATION1 gene mediates the acceleration of flowering by prolonged cold (vernalization) in temperate cereals. This study examined the earliest stages of the transcriptional response of VRN1 to low temperatures. Time-course analyses, using a sensitive quantitative PCR assay, showed that in sprouting barley seedlings VRN1 transcripts begin to accumulate within 24 hours of the onset of cold. The kinetics of the initial transcriptional response of VRN1 to cold was similar to the cold-induced genes DEHYDRIN5 (DHN5) and COLD REGULATED 14B (COR14B), but occurred at lower levels compared to cold acclimation genes or the response to longer cold treatments. Temperatures between 15 and -2 °C induced expression of VRN1 within 24 hours, with a maximal response observed between 2 and -2 °C. Transcriptional induction was also observed in undifferentiated callus cells. There were significant increases in histone acetylation levels at the VRN1 locus in response to 24-hour cold treatment. Sodium butyrate, a histone deacetylation inhibitor, triggered an increase in histone acetylation at VRN1 chromatin and elevated VRN1 transcript levels. The transcriptional response of VRN1 to short-term cold treatment was examined in near-isogenic lines that have different VRN1 genotypes, showing that an allele of the barley VRN1 gene with an insertion in the first intron and high basal expression levels has a reduced transcriptional response to short term cold treatment. This study suggests that low-temperature induction of VRN1 is a cellular response to cold triggered by the same mechanisms that mediate low-temperature induction of cold acclimation genes.

Introduction

Many plants must experience prolonged cold during winter (vernalization) in order to flower rapidly in spring (Chouard, 1960). This coordinates floral development with the seasons by delaying progression towards frost-sensitive reproductive phases of growth until after winter has passed (Mahfoozi ; Limin and Fowler, 2006). Genes controlling vernalization-induced flowering have been identified in the model plant Arabidopsis, in cultivated sugar beet, and the cereal crops wheat and barley (Michaels and Amasino, 1999; Sheldon ; Danyluk ; Trevaskis ; Yan ; Pin ). Different genes mediate vernalization-induced flowering in each of these plant lineages, suggesting that the vernalization response has evolved independently on multiple occasions (Greenup ).

The VERNALIZATION1 gene (VRN1) controls vernalization-induced flowering in cereals and related temperate grasses (Danyluk ; Murai ; Trevaskis , 2007a; Yan ; Preston and Kellogg, 2008; Seppänen ; Asp ; Ergon ). VRN1 encodes a MADS box transcription factor that promotes flowering (Danyluk ; Trevaskis ; Yan ). VRN1 is expressed at low basal levels but transcript levels increase during prolonged cold treatment (Danyluk ; Trevaskis ; Yan ). Activation of VRN1 is quantitative, with longer cold treatments inducing higher levels of expression (Danyluk ; Yan ; von Zitzewitz ; Sasani ). Expression of VRN1 is maintained at elevated levels when vernalized plants return to normal growth temperatures and this promotes flowering (Yan ; Trevaskis ; Hemming ; Sasani ). Elevated VRN1 expression is associated with increased expression of floral promoters, such as FLOWERING LOCUS-T like-1 and FLOWERING PROMOTING FACTOR1-like genes, and decreased expression of floral repressors, such as VERNALIZATION2 (VRN2) and ODDSOC2, suggesting that VRN1 promotes flowering of vernalized plants by regulating the activity of a series of other flowering-time genes, directly or indirectly (Yan , 2006; Hemming ; Sasani ; Shimada ; Greenup , 2011, Chen and Dubcovsky, 2012).

Some alleles of VRN1 have high basal transcript levels and so allow rapid flowering without vernalization (Danyluk ; Murai ; Trevaskis ; Yan ). These alleles have been used to breed varieties of wheat and barley that can be grown where vernalization does not occur, in warm regions of Australia for example (Eagles , 2011). A number of mutations have been identified in alleles of VRN1 that have high basal expression. These include small insertions/deletions or single nucleotide polymorphisms within the proximal promoter and large insertions/deletions within the first intron of the VRN1 locus (Yan , 2004b; Fu ; von Zitzewitz ; Cockram ; Szucs ; Hemming ). These mutations potentially define regions within the VRN1 gene that control transcriptional activity. The repressive histone marker histone 3 lysine 27 tri-methylation occurs at the promoter and first intron of the VRN1 locus (Oliver ). These regions of the VRN1 gene might be targeted by the plant Polycomb Repressor Complex 2 to maintain the chromatin of VRN1 in a repressed state until plants are vernalized (Oliver ).

Two genes, VRN2 and VEGETATIVE TO REPRODUCTIVE TRANSITION 2 (VRT2), have been suggested to mediate low-temperature induction of VRN1. Both were suggested to operate as repressors of VRN1 that are themselves downregulated by cold during winter to allow increased expression of VRN1 to trigger flowering in spring (Yan ; Kane ). Subsequent studies showed that, although both genes are likely to regulate flowering or floral development, neither is likely to mediate low-temperature induction of VRN1. VRN2 is not active in the short days of winter, when low-temperature induction of VRN1 occurs (Dubcovsky ; Trevaskis ; Hemming ; Sasani ). VRT2 activity increases during vernalization, contrary to the original report that it is repressed by cold, which is inconsistent with a role as a repressor that is inactivated by vernalization (Trevaskis ).

Understanding transcriptional regulation of VRN1 is the key to understanding the molecular basis for vernalization-induced flowering in temperate cereals (Trevaskis, 2010). This study examines the earliest stages of low-temperature induction of VRN1 in barley; a model organism for temperate cereals and an important crop. Data are presented showing that VRN1 transcripts accumulate within 24 hours when barley plants are exposed to low temperatures and that this is associated with changes in chromatin state at the VRN1 locus.

Materials and methods

Growth conditions

Seeds were imbibed on 90-mm filter paper discs (Whatman) in Petri dishes with 5ml of water and 1.4g l–1 Thiram fungicide (Bayer Crop Science), then germinated in darkness for 5 days at 20 °C, reaching an average coleoptile length of 4cm, before being placed at 4 °C for 24 hours or maintained at 20 °C for the same period (control). Both cold and control treatments were in darkness, and seedlings were harvested directly into liquid nitrogen at the end of treatment. Sodium butyrate treatment was performed by placing 5ml sodium butyrate solution on filter paper discs once seedlings reached an average coleoptile length of 4cm. For comparisons of different low-temperature treatments, seedlings were germinated in 15ml falcon tubes in perlite/vermiculite (50:50). Tubes were incubated at 20 °C in an Echotherm programmable temperature block (Torrey Pines Scientific Instruments, California, USA), in darkness for 5 days (average coleoptile length of 4cm) then shifted to low temperatures (–2 to 15 °C) for 24 hours or maintained at 20 °C as a control, before harvest.

Plant material

The vernalization-responsive winter barley cultivar Sonja (HvVRN1, HvVRN2, PPD-H1) has been described previously (Sasani ). Near-isogenic lines were developed by five rounds of recurrent crossing to the facultative barley WI4441 (HvVRN1, ΔHvVRN2, PPD-H1). Allele donors were: AUS40413 (HvVRN1-1), AUS405184 (HvVRN1-3), AUS403647 (HvVRN1-4), AUS401872 (HvVRN1-7), AUS403467 (HvVRN1-9), according to the allele notation of Hemming . Near-isogenic lines of wheat were generated by five rounds of recurrent crossing to the Australian spring wheat cv. Sunstate, which carries a wild-type VRN-A1 allele with VRN-B1a and VRN-D1a intron-deletion alleles, and the PPD-D1a allele that confers photoperiod insensitivity. Some of these lines have been described previously (Alonso-Peral ). Allele donors were AUS90785 (VRN-A1a), AUS15490 (VRN-A1b), and AUS7374 (VRN-A1 Langdon). Callus tissue was generated by placing surface sterilised embryos of cv. Golden Promise on callus induction media (Tingay ) for 4 weeks in darkness. Calli were then maintained on tissue culture media at 20 °C or shifted to 4 °C for 24 hours and then harvested for gene expression analysis.

Gene expression analyses

RNA was extracted using the method of Chang . Total RNA (5 μg) was reverse-transcribed with Super Script III reverse transcriptase (Invitrogen), according to manufacturer instructions. Quantitative real-time PCR was performed in a Rotorgene Q real-time PCR machine. Expression was normalized to ACTIN using the Rotorgene software (Qiagen), which takes amplification efficiency into account. Primers for ACTIN and the wheat VRN-A1 homeoallele have been described previously (Trevaskis ; Alonso-Peral ). Primers used for other genes are presented in Supplementary Table S1 (available at JXB online). Comparison of HvVRN1 expression between different near-isogenic lines was performed on a 7900HT Fast Real-Time PCR System (Applied Biosystems) with SYBR green and Platinum DNA polymerase (Invitrogen). ACTIN was used as reference and relative transcript levels of biosynthesis were calculated with the ΔΔCt method, factoring in primer amplification efficiencies.

Chromatin immunoprecipitation

Nuclei extraction and immunoprecipitation was performed as described previously (Oliver ). The antibody against histone H3 was purchased from Abcam, against acetylated histone 3 (H3Ac, lysine 9 (K9) and K14) and acetylated histone 4 (H4Ac, K5, K8, K12, K16) from Millipore. The amount of genomic DNA precipitated in ChIP assays was quantified by quantitative real-time PCR. Primer pairs used for HvVRN1 were: 5’-ACCCCAAGTGGAAGGGTTAG-3’, 5’-TTTGTTCAAG CAGCAAGCAC-3’ (2708–2872bp relative to the HvVRN1 allele of cv. Strider (Genbank AY750993) region 3 of Fig. 1A in Oliver ). Primer pairs used for ACTIN have been described previously (Oliver ). For each primer pair, the amount of DNA precipitated using anti-H3Ac or anti-H4Ac antibodies was normalized to the amount precipitated by an anti-H3 antibody from the same sample to correct for differences in amounts of ChIP input DNA. ACTIN (enriched for H3Ac and H4Ac) was used for normalization to compare cold-treated and control samples (expression of this gene does not change with cold). No-antibody control reactions were performed in parallel with each antibody reaction to verify that the precipitated DNA was enriched for ACTIN for the H3Ac and H4AC immunoprecipitations, and for ACTIN and HvVRN1 for the H3 immunoprecipitations). The data presented is the relative amount of precipitated DNA normalized to ACTIN, and each graph represents the average of two technical repeats from a minimum of two separate experiments.

Fig. 1.

Kinetics of HvVRN1, DHN5, and COR14b transcript accumulation in barley seedlings after the shift to cold. Transcript levels of HvVRN1 (A) DHN5 (B), or COR14b (C), relative to ACTIN, in seedlings (cv. Sonja) maintained at 20 °C in darkness (dashed lines with solid squares) or at various time points after being shifted to 4 °C in darkness (solid line with open circles). Error bars show standard error of a minimum of three biological repeats. * P < 0.05, ** P < 0.01, *** P < 0.001, Student’s t-test, control versus treatment at time point indicated.

Results

HvVRN1 transcript levels begin to increase within 24 hours of the onset of low temperatures

Transcript levels of the barley VRN1 gene (HvVRN1) were assayed to determine when expression first begins to increase in response to low temperatures. In seedlings of the vernalization-responsive barley cultivar Sonja, which has a wild-type version of HvVRN1 with the full-length first intron, a 2-fold increase in HvVRN1 transcript levels was detected after 12 hours of cold treatment, with further increased transcript levels after 24 hours (Fig. 1A). This expression pattern is similar to that of DEHYDRIN5 (DHN5) and COLD REGULATED 14B (COR14B) (Fig. 1B,C), two genes shown by transcriptome analyses to be useful markers for transcriptional responses to low temperatures in barley (Tommasini ; Greenup ). The levels of HvVRN1 transcript detected after 24 hours of cold were low compared to prolonged cold treatments, unlike cold acclimation genes which were induced to high levels within a single day of cold treatment and remained high with continuing cold treatment (Supplementary Fig. S1).

Induction of C-repeat transcription factors precedes cold induction of HvVRN1

C-REPEAT BINDING FACTORS (CBF) genes are rapidly upregulated when barley seedlings are exposed to low temperatures and are thought to activate cold acclimation pathways (Choi ; Xue, 2003; Skinner ; Stockinger ). The transcript levels of five CBF transcription factor genes were assayed in barley seedlings during the first 24 hours of cold treatment. Expression of HvCBF1, HvCBF2, and HvCBF6 increased rapidly (within 3 hours) but returned to basal expression levels within 6 hours (Fig. 2A,B). Expression of HvCBF4 also increased rapidly but showed a slower decrease in activity, returning to basal expression levels by 9 hours (Fig. 2D). HvCBF9 transcript levels increased rapidly, peaking within 3 hours, but remained elevated throughout the subsequent time points, unlike the other CBF genes assayed (Fig. 2E).

Fig. 2.

Kinetics of CBF transcript accumulation in barley seedlings after the shift to cold. Transcript levels of HvCBF1 (A), HvCBF2 (B), HvCBF4 (C), HvCBF6 (D), or HvCBF9 (E). Expression is shown relative to ACTIN, assayed in seedlings (cv. Sonja) maintained at 20 °C (dashed lines with solid squares) or at various time points after being moved to 4 °C (solid line with open circles). Both temperature treatments were in darkness. Error bars show standard error of a minimum of three biological repeats. * P < 0.05, ** P < 0.01, *** P < 0.001, Student’s t-test, control versus treatment at time point indicated.

Transcription of HvVRN1 is induced to a greater extent as temperatures approach freezing

Seedlings were shifted from warm conditions (20 °C) to a range of temperatures from 15 to –2 °C for 24 hours. Induction of HvVRN1 occurred in all treatments, but to different extents (Fig. 3A). Mild chilling (15 or 10 °C) induced a low level of HvVRN1 expression, with transcript levels induced to higher levels as temperatures approached 0 °C. In comparison, expression of COR14B showed a similar induction optimum of between 2 and –2 °C, whereas DHN5 was induced to the greatest extent at 2 °C (Fig. 3B,C).

Fig. 3.

Transcript levels of HvVRN1, DHN5, and COR14b after 24 hours at different temperatures. Transcript levels of HvVRN1 (A) DHN5 (B), or COR14b (C), relative to ACTIN, in seedlings (cv. Sonja) maintained at 20 °C or after being moved to temperatures between 15 and –2 °C for 24 hours. All temperature treatments were applied in darkness. Error bars show standard error of a minimum of three biological repeats. * P < 0.05, ** P < 0.01, *** P < 0.001, Student’s t-test, control versus treatment. ND, expression of VRN1 was not detected.

Low-temperature induction of HvVRN1 occurs in undifferentiated cells

Undifferentiated callus tissue can be generated from barley embryos of the spring barley cv. Golden Promise by tissue culture. This barley has an allele of HvVRN1 with a 5.8-kb deletion in the first intron that has high basal expression levels. Nevertheless there was a significant increase in HvVRN1 transcript levels in seedlings of this cultivar following 24-hour cold treatment (Fig. 4A, Supplementary Fig. S2). Transcript levels of HvVRN1 also increased in callus cells subjected to 24-hour cold (Fig. 4B). Similarly, expression of DHN5 increased after 24-hour cold treatment in callus (Fig. 4C), but there was no increase in COR14b expression in callus of this barley cultivar (Fig. 4C).

Fig. 4.

Transcript levels of HvVRN1 in Golden Promise seedlings or callus tissue after 24-hour cold treatment. (A) Transcript levels of HvVRN1 in seedlings of the spring barley cv. Golden Promise at 20 °C (control) versus seedlings shifted from 20 to 4 °C for 24 hours (cold). (B) Transcript levels of HvVRN1 in 1-cm diameter callus generated from embryos of (cv. Golden Promise) at 20 °C (control) versus callus shifted from 20 to 4 °C for 24 hours (cold). (C) Transcript levels of DHN5 in control versus 24-hour cold-treated callus. (D) Transcript levels of COR14b in control versus 24-hour cold-treated callus. Expression is shown relative to ACTIN. Error bars show standard error of a minimum of three biological repeats. * P < 0.05, Student’s t-test, control versus treatment. Both temperature treatments were in darkness.

Low temperatures trigger rapid changes in histone acetylation levels at the HvVRN1 locus

The effect of short-term cold on the state of chromatin in the first intron of the HvVRN1 locus was analysed by measuring histone acetylation levels, a marker of active gene transcription. Levels of acetylation at histones H3 and H4 increased in HvVRN1 chromatin within 24 hours of the onset of cold (Fig. 5A,B) and continued to increase with longer cold treatments (Fig. 5C). Seedlings were then treated with sodium butyrate an inhibitor of histone deacetylase activity. After 24-hour incubation in sodium butyrate, the levels of acetylation at histones H3 and H4 were increased in chromatin in the first intron of HvVRN1 compared to the control treatment (Fig. 6A,B). HvVRN1 transcript levels increased in response to sodium butyrate treatment (Fig. 6C).

Fig. 5.

Histone acetylation (H3Ac) levels increase at HvVRN1 chromatin during cold treatment. (A) H3Ac levels in HvVRN1 chromatin from control or cold-treated seedlings (24 hours at 4 °C, cv. Sonja). (B) H4Ac levels in HvVRN1 chromatin from control (20 °C) or 24-hour cold-treated (4 °C) seedlings. (C) H3Ac levels in HvVRN1 chromatin during different durations of cold treatment from 1 to 28 days. * P < 0.05, ** P < 0.01, *** P < 0.001, Student’s t-test, control versus treatment. All treatments were in darkness.

Fig. 6.

Sodium butyrate triggers increased histone acetylation (H3Ac) at HvVRN1 chromatin and increased transcript levels. (A) H3Ac levels in HvVRN1 chromatin from control versus sodium butyrate treated seedlings (25mM butyrate, 24 hours). (B) H4Ac levels in HvVRN1 chromatin from control versus sodium butyrate treated seedlings. Histone acetylation levels are normalized to ACTIN. Error bars show standard error. (C) HvVRN1 expression in seedlings exposed to different concentrations of sodium butyrate for 24 hours, compared to untreated controls. Expression is normalized to ACTIN, error bars show standard error. * P < 0.05, ** P < 0.01, *** P < 0.001, Student’s t-test, control versus treatment. All treatments were in darkness.

The HvVRN1-7 allele with an insertion in the first intron does not show a transcriptional response to 24-hour cold treatment

Expression of HvVRN1 was assayed in near-isogenic lines of barley that carry alleles with different insertions or deletions within the first intron of the HvVRN1 gene. The levels of HvVRN1 transcripts varied between seedlings of the different lines in control conditions (Table 1, Supplementary Fig. S3). The parent line with a wild-type HvVRN1 allele (full-length first intron and no proximal promoter mutations) had the lowest HvVRN1 transcript levels whereas a line with the HvVRN1-7 allele, with a 0.7-kb insertion in the first intron (first identified in cv. Varunda, Cockram ), had the highest levels, approximately 300-fold higher than those associated with the wild-type allele (Table 1). HvVRN1 alleles with intron deletions had a range of expression levels (Table 1, Supplementary Fig. S3). Generally, the near-isogenic lines showed elevated HvVRN1 expression following 24-hour cold treatment (Table 1). The only exception was the line carrying the HvVRN1-7 allele, where there were no differences in HvVRN1 transcript levels between cold versus control treatments (Table 1, Supplementary Fig. S4).

Table 1.

Expression of HvVRN1 in near-isogenic barley lines. Basal: expression ratio of different barley VRN1 alleles versus the wild-type HvVRN1 allele, in seedlings at 20 °C in darkness. Cold: expression ratio for 24-hour cold-treated versus control seedlings (fold-change expression with cold treatment, in darkness). P1: P-values for Student’s t-test for basal expression comparison with each VRN1 allele versus the wild type. P2: P-values for VRN1 expression in control versus 24-hour cold-treated seedlings. Descriptions of different VRN1 alleles of barley, and size of insertions of deletions in the first intron, as compared to the wild-type allele sequence from the winter barley cv. Strider (accession number AY750993). Genbank accessions of other alleles: HvVRN1-9, FJ687749; HvVRN1-4, DQ492704; HvVRN1-1, AY750995; HvVRN1-3, EF591642; HvVRN1-7, EF591650.

Allele	Basal	P1	Cold	P2	Description	Size
Wild type	1	NA	1.8	7.4E–03	Full-length first intron	NA
HvVRN1-9	6	4.0E–03	2.6	9.3E–05	Intron deletion	2.4
HvVRN1-4	9	6.7E–04	2.9	1.1E–04	Intron deletion (cv. Calu. Sib)	4.1
HvVRN1-1	35	9.4E–04	2.1	9.3E–05	Intron deletion (cv. Morex)	5.1
HvVRN1-3	70	6.9E–08	2.1	9.7E–06	Intron deletion (cv. Triumph)	8.7
HvVRN1-7	312	2.1E–03	1.0	NS	Intron insertion (cv. Varunda)	0.7

Mutations in the promoter of the hexaploid wheat A genome copy of VRN1 do not abolish rapid transcriptional induction by cold

Expression of VRN1 was assayed in near-isogenic lines of wheat that carry different alleles of the VRN1 gene on the A genome (VRN-A1). These alleles showed different basal expression levels in control conditions (Table 2, Supplementary Fig. S5), with an insertion in the promoter (VRN-A1a, Yan ) associated with the highest basal transcript levels. An allele with a point mutation plus a small deletion in the promoter (VRN-A1b, Yan ) and an allele with a large deletion in the first intron (same as in the durum wheat cv. Langdon, Fu ) had intermediate basal transcript levels. All alleles showed elevated transcript levels following 24-hour cold treatment (Table 2).

Table 2.

Expression of VRN-A1 in near-isogenic wheat lines. Basal: expression ratio of different wheat VRN1 alleles versus the wild-type VRN1-A1 allele, in seedlings at 20 °C in darkness. Cold: expression ratio for 24-hour cold-treated versus control seedlings. P1: P-values for Student’s t-test for basal expression comparison with each VRN1 allele versus the wild type. P2: P-values for VRN1 expression in control versus 24-hour cold-treated seedlings. Descriptions of different VRN-A1 alleles (VRN1 gene on the A genome) of wheat, as compared to the wild-type allele sequence from the winter wheat test line Triple Dirk C. Genbank accessions of alleles: VRN-A1, AY747600; VRN-A1a, AY747601. VRN-A1 Langdon, AY747598; VRN-A1b, AY616461.

Allele	Basal	P1	Cold	P2	Description
VRN-A1	1	NA	6.3	1.7E–04	Wild type
A1lang	10.5	1.7E–07	1.8	8.2E–04	Intron deletion
A1b	15.0	6.0E–07	3.7	5.0E–10	Promoter single-nucleotide polymorphism, 20-bp deletion
A1a	75.1	1.7E–10	4.1	8.2E–03	231-bp promoter insertion

Discussion

The data presented show that transcripts of HvVRN1 start to accumulate rapidly after the onset of cold, with a 2-fold increase within 12 hours, and an 8-fold increase by 24 hours (Fig. 1). These data are consistent with a previous study that reported induction of VRN1 after 24-hour cold treatment in a winter wheat using a semi-quantitative assay, but did not examine earlier time points (Kobayashi ). The level of VRN1 expression after 24 hours of cold is low compared to that observed at the end of longer cold treatment (Supplementary Fig. S1). Unlike longer cold treatments (Sasani ), elevated expression of HvVRN1 was not maintained in the leaves of seedlings that were returned to normal glasshouse conditions (Supplementary Fig. S1). Nor was there any visible acceleration of shoot apex development by a single day of cold (Supplementary Fig. S6). Thus, the initial induction of HvVRN1 in seedlings exposed to a single day of cold is the first stage of a gene expression response and is important in terms of understanding the mechanisms that activate expression of VRN1 at low temperatures, but is not sufficient for flowering response.

Although the initial transcriptional response of HvVRN1 to cold was weak compared to extended cold treatments, there was a significant increase in the level of histone acetylation at the HvVRN1 locus after 24-hour cold (Fig. 5). It is possible that histone acetylation is an early step in the process that activates transcription of VRN1 during vernalization, possibly promoting a shift towards an active chromatin state. As was the case for transcript levels, the magnitude of chromatin modifications increased with longer cold treatments suggesting that the initial changes observed after a single day of cold were the beginning of a process that continues with increasing duration of cold (Fig. 5). Treating barley seedlings with sodium butyrate, an inhibitor of histone deacetylation, increased acetylation at the HvVRN1 locus and elevated HvVRN1 transcript levels, supporting a role for histone acetylation/deacetylation in regulating VRN1 expression (Fig. 6).

The increased levels of histone acetylation at the HvVRN1 locus and increased HvVRN1 transcript levels detected within 24 hours of the onset of low temperatures supports the hypothesis that induction of VRN1, and by extension vernalization-induced flowering in cereals, is a specific response to low temperatures. An alternative hypothesis, that vernalization is a response to altered rates of development at low temperatures (Allard ), seems less likely. Induction of HvVRN1 in the seedlings is unlikely to be triggered by developmental effects since development occurs over longer time scales (1 leaf initiated per 5 days at 20 °C) and the experiments described here used developmentally matched samples for gene expression analysis. Furthermore, altered developmental patterns cannot account for the rapid induction of VRN1 in undifferentiated cells (Fig. 4).

Comparison of the expression levels of different alleles of HvVRN1 in near-isogenic lines extended previous studies that examined expression of different alleles in diverse barley cultivars (Hemming ). An allele with an insertion in the first intron is associated with the strongest basal activity (Table 1, Supplementary Fig. S3). Plants with alleles that lack large segments that extend towards the start of the first intron were also associated with a large increase HvVRN1 transcript levels, whereas smaller deletions (HvVRN1-4, HvVRN1-9) are associated with moderate basal transcript levels (Table 1, Supplementary Fig. S3). Genetic data show that alleles with even a modest increase in transcript levels in sprouting seedlings are associated with some degree of early flowering without vernalization (HvVRN1-4: Szucs ; Casao ). In near-isogenic lines of wheat, an insertion in the promoter (VRN-A1a) was associated with the highest basal expression levels, followed by another promoter mutation allele (VRN-A1b) and an intron deletion allele (Table 2, Supplementary Fig. S5). Thus, different VRN1 alleles of both wheat and barley have a wide spectrum of basal activity levels.

Most of the barley and wheat alleles of VRN1 assayed showed significantly increased transcript levels following 24-hour cold treatment, irrespective of basal activity levels. The only VRN1 allele that showed no increase in expression following 24-hour cold treatment was the HvVRN1-7 allele that has a 0.7-kb insertion in the first intron (Supplementary Fig. S4). Previous studies have found that this allele retains a response to longer cold treatments (Stockinger ; Hemming ). The insertion in the first intron might disrupt a regulatory mechanism that acts specifically during the initial response to low temperatures. Alternatively, the initial response to cold might be difficult to detect compared to the high basal expression level of this allele.

The DHN5 and COR14B genes are rapidly activated by cold and suggested to contribute to increased freezing tolerance when plants are exposed to low temperatures (Crosatti ; Danyluk ). The rapid initial response of HvVRN1 to low temperatures parallels that of DHN5 and COR14B, although HvVRN1 is expressed at lower levels (Fig. 1). The weaker transcriptional response of HvVRN1 is likely due to the repression mechanisms that target the first intron of the HvVRN1 gene, which lower basal gene expression levels and add a requirement for prolonged cold to allow the accumulation of high transcript levels. As was the case for HvVRN1, DHN5 showed elevated transcript levels in undifferentiated cells following 24-hour cold treatment. No cold induction of COR14b was observed in callus of the spring barley Golden Promise, but this gene is induced by 24 hours of cold treatment in callus generated from the winter barley cv. Nure, suggesting varietal differences in regulation of this gene (Vashegyi ).

The overall similarities between the expression patterns of DHN5, COR14B, and HvVRN1 during short-term cold treatment suggest that common signalling pathways regulate cold acclimation and the initial stages of the vernalization response. Furthermore, since expression of all three genes increased more when seedlings were exposed to 2 versus 4 °C (Fig. 3), it seems that rapid initial transcriptional induction is triggered by a temperature-sensing pathway (or pathways) that can discriminate small differences in temperature, even at low temperatures.

Induction of CBF transcription factor genes precedes the initial increase in transcript levels of DHN5, COR14B, and HvVRN1 transcripts (Fig. 2). The promoters of all three genes contain potential binding sites for CBF transcription factors (Choi ; Dal Bosco ; Skinner ; Alonso-Peral ) and transgenic barley overexpressing CBF genes show altered expression of DHN5 and COR14b (Morran ). In Arabidopsis, CBF transcription factors interact with histone acetylation complexes (Stockinger ), and histone acetylation has been shown to play an important role in activating expression of cold acclimation genes (Pavangadkar ). Low-temperature-induced changes in histone acetylation levels at the first intron of HvVRN1 (Fig. 5), and rapid increases in transcript levels, could potentially be driven by CBF transcription factors binding to sites within the promoter of HvVRN1.

In summary, the data presented show that the vernalization response of cereals has the following features: (1) a rapid response to low temperatures, influencing both chromatin modifications and transcript levels of VRN1; (2) a low-temperature sensor (or sensors) active in a range of cell types including undifferentiated cells; and (3) a low-temperature sensor (or sensors) sensitive to small differences in temperature, even at temperatures well below those optimal for growth of barley. This study suggests that these are features common to both the vernalization and cold acclimation responses, and that in temperate cereals a common regulatory mechanism (or mechanisms) controls the initial stages of both these responses to the cold of winter.

Supplementary material

Supplementary data are available at JXB online.

Supplementary Table S1. Primers used for gene expression analysis.

Supplementary Fig. S1. Expression of HvVRN1 and DHN5 at various time points during or after vernalization, compared to non-vernalized controls.

Supplementary Fig. S2. Expression of HvVRN1 during short-term cold treatment of seedlings of cv. Golden Promise.

Supplementary Fig. S3. Comparison of basal transcript levels of different HvVRN1 alleles of near-isogenic barley lines.

Supplementary Fig. S4. Comparison of expression of the HvVRN1 and HvVRN1-7 alleles in near-isogenic barley seedlings with or without cold treatment.

Supplementary Fig. S5. Comparison of basal transcript levels of different VRN1 alleles of near-isogenic wheat lines.

Supplementary Fig. S6. Shoot apices from plants at the third leaf stage after 1D cold treatment as seedlings, versus non-treated controls.