Wheat TILLING mutants show that the vernalization gene VRN1 down-regulates the flowering repressor VRN2 in leaves but is not essential for flowering

Abstract

Most of the natural variation in wheat vernalization response is determined by allelic differences in the MADS-box transcription factor VERNALIZATION1 (VRN1). Extended exposures to low temperatures during the winter (vernalization) induce VRN1 expression and promote the transition of the apical meristem to the reproductive phase. In contrast to its Arabidopsis homolog (APETALA1), which is mainly expressed in the apical meristem, VRN1 is also expressed at high levels in the leaves, but its function in this tissue is not well understood. Using tetraploid wheat lines with truncation mutations in the two homoeologous copies of VRN1 (henceforth vrn1-null mutants), we demonstrate that a central role of VRN1 in the leaves is to maintain low transcript levels of the VRN2 flowering repressor after vernalization. Transcript levels of VRN2 were gradually down-regulated during vernalization in both mutant and wild-type genotypes, but were up-regulated after vernalization only in the vrn1-null mutants. The up-regulation of VRN2 delayed flowering by repressing the transcription of FT, a flowering-integrator gene that encodes a mobile protein that is transported from the leaves to the apical meristem to induce flowering. The role of VRN2 in the delayed flowering of the vrn1-null mutant was confirmed using double vrn1-vrn2-null mutants, which flowered two months earlier than the vrn1-null mutants. Both mutants produced normal flowers and seeds demonstrating that VRN1 is not essential for wheat flowering, which contradicts current flowering models. This result does not diminish the importance of VRN1 in the seasonal regulation of wheat flowering. The up-regulation of VRN1 during winter is required to maintain low transcript levels of VRN2, accelerate the induction of FT in the leaves, and regulate a timely flowering in the spring. Our results also demonstrate the existence of redundant wheat flowering genes that may provide new targets for engineering wheat varieties better adapted to changing environments.

Figure 1. Effect of photoperiod and vernalization on wheat flowering time.

During the fall, VRN2 competes successfully with CO (photoperiod pathway, FT promoter) for interactions with the NF-Y transcription factors, resulting in the down-regulation of FT transcription in the leaves . This precludes flowering in the fall. Vernalization induces VRN1 and down-regulates VRN2 transcription in the leaves. The presence of VRN1 after the winter is important to maintain the down-regulation of VRN2 during the spring. In the absence of VRN2, FT transcription is up-regulated and the encoded FT protein is transported through the phloem to the stem apical meristem. FT then interacts with FDL2 to up-regulate VRN1 transcripts to the levels required for the transition to the reproductive phase. Dashed red lines indicate interactions demonstrated in this study.

Figure 2. Heading times and spikelet morphology of VRN1 mutants.

Δvrn-A1: mutation in VRN-A1 (winter, functional vernalization responsive VRN-B1 allele, functional VRN2), Δvrn-B1: mutation in VRN-B1 (spring, functional vernalization insensitive VRN-A1 allele, functional VRN2), Δvrn1-null: truncated VRN1 proteins (functional VRN2, winter), Δvrn1-Δvrn2-null: no functional VRN1 and VRN2 proteins (early). A) Growth chamber experiment: Heading times of unvernalized and 6-weeks vernalized Δvrn-A1 and Δvrn1-null mutants set 1 (premature stop codon, top) and 2 (splice site mutant, bottom). Control = Δvrn-B1 (functional vernalization-insensitive Vrn-A1 allele). B) Top: maturity differences between wild type, single and null VRN1 mutants (set1). Bottom: spikelet morphology at anthesis and mature seeds from wild type and Δvrn1-null mutants. C) Greenhouse experiment comparing heading times of unvernalized and 8-weeks vernalized Δvrn1-null and Δvrn1-Δvrn2-null mutants. These sib lines have the same VRN1 mutations but differ in the presence or absence of functional VRN2 genes. Control = Δvrn-B1-Δvrn2-null (functional vernalization-insensitive VRN-A1 allele and no functional VRN2 genes). Heading times of the vernalized lines (black bars) are adjusted using the difference in flowering time between vernalized and unvernalized spring control lines as described in Material and Methods.

Figure 3. qRT–PCR transcriptional profiles of VRN1, FT, and ZCCT2 ( = VRN2) in mutant set 1 (premature stop codon).

A–B) VRN1, C–D) FT, E–F) ZCCT2. Left panels A, C and E) unvernalized plants. Right panels B, D, and F) vernalized plants. Δvrn-A1: mutation in VRN-A1 (winter, functional vernalization responsive VRN-B1 allele), Δvrn-B1: mutation in VRN-B1 (spring, functional vernalization insensitive VRN-A1 allele), Δvrn1-null: truncated VRN1 proteins. Blue shaded areas indicate vernalization at 4°C under long days. 0 wV: 3 weeks-old plants grown under 22°C/17°C (day/night) conditions before vernalization, 3 wV: 3 weeks of vernalization, 6 wV: 6 weeks of vernalization, RT: two weeks after returning the vernalized plants to pre-vernalization conditions. A final sample was obtained from flag leaves at heading times (FL, dotted lines), which are indicated in days from sowing to heading. The X axis scale is not proportional to time and the Y scale is in fold-ACTIN values. Error bars are SE of the means from 8 biological replications.

Figure 4. Comparison of VRN2 and FT transcript levels between Δvrn1-null and Δvrn1-Δvrn2-null mutants.

A–B) ZCCT2, C–D) FT. Left panels A and C) unvernalized plants. Right panels B and D) vernalized plants. Δvrn1-null: no functional VRN1 proteins (functional VRN2, winter), Δvrn1-Δvrn2-null: truncated VRN1 and VRN2 proteins (early). Blue shaded areas indicate vernalization at 4°C under long days. 0 wV: 3 weeks-old plants grown under 22°C/17°C (day/night) conditions before vernalization, 4 wV: 4 weeks of vernalization, 8 wV: 8 weeks of vernalization, RT: two weeks after returning the vernalized plants to pre-vernalization conditions. A–B) ZCCT2 ( = VRN2). Lower transcript levels in the Δvrn1-Δvrn2-null mutant (red line) are likely caused by the complete deletion of the VRN-B2 genes in this line. Only the non-functional ZCCT-A2 transcripts are detected, C–D = FT. Note the rapid up-regulation of FT in the Δvrn1-Δvrn2-null mutants relative to the Δvrn1-null mutants. The X axis scale is in weeks (w) and is not proportional to time. The Y scale is in fold-ACTIN values. Error bars are SE of the means from 8 biological replications.

Figure 5. Transcriptional profiles of FUL2, FUL3, and FT during and after vernalization in the Δvrn1-null mutant set 1 (premature stop codon).

A–B) FUL3, C–D) FUL2, and FT. Left panels A and C) samples from leaves. Right panels B and D) samples from apical region. The blue shaded area indicates vernalization at 4°C under long days. 0 wV: 3 weeks-old plants grown at 22°C/17°C (day/night) immediately before vernalization, 3 wV: 3 weeks of vernalization, 6 wV: 6 weeks of vernalization, RT: two weeks after removing the plants from the cold and returning them to pre-vernalization conditions. The X axis scale is not proportional to time and the Y scale is in fold-ACTIN values. Error bars are SE of the means from 8 biological replications in the leaves and 3 biological replications in the apices (each including a pool of 30 shoot apical meristem and surrounding tissue).

Figure 6. Alternative flowering models in the temperate grasses.

A) ‘Reverse model’ , B) ‘Original model’ , .Green arrows indicate promotion of transcription and red lines repression of transcription.