The Arabidopsis mutant, fy-1, has an ABA-insensitive germination phenotype.

Arabidopsis FY, a homologue of the yeast RNA 3' processing factor Pfs2p, regulates the autonomous floral transition pathway through its interaction with FCA, an RNA binding protein. It is demonstrated here that FY also influences seed dormancy. Freshly-harvested seed of the Arabidopsis fy-1 mutant germinated readily in the absence of stratification or after-ripening. Furthermore, the fy-1 mutant showed less ABA sensitivity compared with the wild type, Ler, under identical conditions. Freshly-harvested seed of fy-1 had significantly higher ABA levels than Ler, even though Ler was dormant and fy-1 germinated readily. The PPLPP domains of FY, which are required for flowering control, were not essential for the ABA-influenced repression of germination. FLC expression analysis in seeds of different genotypes suggested that the effect of FY on dormancy may not be elicited through FLC. No significant differences in CYP707A1, CYP707A2, NCED9, ABI3, and ABI4 were observed between freshly-harvested Ler and fy-1 imbibed for 48 h. GA3ox1 and GA3ox2 rapidly increased over the 48 h imbibition period for fy-1, with no significant increases in these transcripts for Ler. ABI5 levels were significantly lower in fy-1 over the 48 h imbibition period. The results suggest that FY is involved in the development of dormancy and ABA sensitivity in Arabidopsis seed.

Introduction

A highly conserved eukaryotic gene, FY, was first identified in one of the late-flowering mutants of Arabidopsis thaliana and shown to be involved in the autonomous flowering pathway (Koornneef ). FY shares homology to a yeast 3′ end-polyadenylation factor Pfs2p, required for maintaining transcript stability by polyadenylation of the 3′ ends of the nascent transcripts (Ohnacker ). The FY gene is comprised of 18 exons and 17 introns and is localized on chromosome V in Arabidopsis. The encoded protein of 84 kDa contains seven WD repeats and two Pro-Pro-Leu-Pro-Pro (PPLPP) domains in the C-terminal region (Simpson ). The WD region is highly conserved among different species, whereas the PPLPP region is relatively less conserved.

The FY protein was first purified from cauliflower extracts through its interaction with FCA, an RNA binding protein involved in regulating flowering in Arabidopsis. This interaction occurred through the WW domain present on FCA and the PPLPP domains present on FY (Simpson ). FCA negatively autoregulates its expression by modulating the site of 3′ polyadenylation in its own pre-mRNA, leading to the formation of four different transcripts (Quesada ). A lack of negative autoregulation of FCA in fy-1 mutants, that lack the PPLPP domains, suggested that FY functions as a 3′-end processing factor in Arabidopsis (Simpson ). FCA-FY interaction leads to the repression of Flowering Locus C (FLC), a MADS-box transcription factor that is a potent repressor of flowering (Michaels and Amasino, 1999; Sheldon ). FCA and FY can operate independently of each other to affect the regulation of FLC through RNA-mediated chromatin silencing (Manzano ).

FY may have pleiotropic effects within the autonomous pathway, as suggested by genetic analysis (Koornneef ). FY may undertake various functions depending on its interacting partners (Henderson ; Manzano ), although no specific functions have been identified. The FY mutants, fy-1and fy-2, both lack PPLPP motifs but retain the WD repeats intact. These mutants are late flowering due to the lack of FCA interaction, resulting in continued FLC expression. The interaction of FCA with FY which was thought to be strong and stable, is now considered transient, leading to altered organization of splicing machinery on FY (Manzano ). Studies with various fy mutants have shown that FY functions in flowering and embryo development and may have phenotypic effects dependent or independent of FCA (Koornneef ; Henderson ; Manzano ).

In addition to flowering, FLC also appears to regulate temperature-dependent seed germination in Arabidopsis (Helliwell ; Chiang ). This is the only documented evidence of any of the components of the autonomous flowering pathway functioning in germination. Increased germination at cool, but not warm, temperatures may be caused by highly expressed natural FLC alleles in Arabidopsis that result in increased expression of CYP707A2 (Chiang ), a gene encoding an ABA 8'-hydroxylase (Kushiro ). ABA 8'-hydroxylase genes have been purported to play a pivotal role in regulating seed dormancy (Millar ), metabolizing ABA and relieving the block on germination. ABA is known to be a factor inhibiting both endosperm rupture and radicle growth (Finch-Savage and Leubner-Metzger, 2006), with the effect being regulated through several transcription factors identified by examining ABA-insensitive mutants in Arabidopsis (Koornneef ). One of these, ABI5, is a basic leucine zipper transcription factor that regulates a number of ABA-inducible responses, such as certain late embryogenesis-abundant genes (Finkelstein and Lynch, 2000; Lopez-Molina and Chua, 2000), embryo lipid mobilization and repression of seed germination in the endosperm (Penfield ). ABI3 and ABI5 are expressed in both Arabidopsis endosperm and embryo, with ABI5 expression confined to the micropylar region of the endosperm (Penfield ). ABI4 is expressed only in the embryo. Piskurewicz suggest that ABI5 acts as the final common repressor of germination in response to changes in ABA and GA levels. They note, however, that expression of ABI5 alone was not sufficient to repress germination as ABI5 phosphorylation is required for activity.

In this work, evidence is provided to demonstrate that another autonomous flowering pathway gene, FY, is also involved in regulating germination. Our experiments show that the Arabidopsis fy-1 mutant lacks seed dormancy, characterized by insensitivity to ABA. The results suggest that reduced levels of ABI5 in Arabidopsis fy-1 may be linked to this ABA-insensitivity.

Materials and methods

Plant materials and growth conditions

Arabidopsis fy-1 mutant (CS57) seeds and its wild-type background parent (Ler) were obtained from the Arabidopsis Biological Resource Centre (ABRC, http://www.arabidopsis.org). Identification of the fy-1 mutant by genomic PCR and sequencing was according to Simpson . Arabidopsis seeds were surfaced-sterilized in 50% (v/v) bleach before sowing on Murashige and Skoog (MS) medium with 1% (w/v) sucrose, stratified for 4 d at 4 °C, and kept in a tissue culture cabinet (22 °C, 16 h photoperiod) for germination. Seven-day-old seedlings were transferred to soil and grown in growth chamber under constant conditions (20 °C, 70% relative humidity, 16 h photoperiod, light intensity of 100 μE m−2 s−1−. All genotypes were grown side by side in a growth chamber. The seeds were harvested and used for further studies.

Seed germination assays

Mature seeds of different genotypes were harvested at the same time from dry siliques (moisture content of the harvested seeds was approximately 9%). Fresh seed denotes seed that was sown directly after harvesting. To afterripen, harvested seed was stored at 22 °C for 4 weeks in darkness. Cold stratification was performed by incubating the imbibed seeds at 4 °C for 3 d in darkness. Batches of 40–70 seeds from each genotype were sown in triplicate in Petri dishes with two layers of water-saturated filter paper (Whatman) or were supplemented with different concentration of ABA (Sigma-Aldrich). Germination was scored daily by radicle emergence for up to 5 d, and the germination frequency (%) was calculated. Each germination experiment was repeated at least three times using biological replicates.

RNA extraction, Northern blots, and real-time RT-PCR

The extraction of total RNA from seeds and Northern blots were performed as previously described by Baumbusch . The primers used for amplifying the probe for Northern blots analysis were as follow: FY, forward 5′-ATGTACGCCGGCGGCGATATG-3′ and reverse 5′-TCCGAATAGGCAACTGTGGG-3′; FLC, forward 5′-GTAGCCGACAAGTCACCTTCTC-3′ and reverse 5′-CTAATTAAGTAGTGGGAGAGTCAC-3′; ACTIN , forward 5′-GAYTCTGGWGATGGTGTG-3′ and reverse 5′-ACCTGYTGKAAHGTRCTGAG-3′. For the real-time RT-PCR, the extracted total RNA was treated with DNaseI (Qiagen, http://www1.qiagen.com) for digestion of genomic DNA, and was further purified using the RNeasy plant mini kit (Qiagen, http://www1.qiagen.com) according to the manufacturer’s instructions. First-strand cDNA was synthesized with 5 μg of total RNA in 50 μl reaction by M-MLV Reverse Transcriptase (Promega, http://www.progema.com). Quantitative RT-PCR was performed, in triplicate, with 1 μl of cDNA template in 25 μl reaction using QuantiTect® SYBR Green PCR kit (Qiagen, http://www1.qiagen.com) and run on the Applied Biosystem 7500 Fast real-time PCR system (http://www.appliedbiosystems.com) according to the manufacturer’s instructions. A melting curve was used to verify the specificity of the amplified fragment. ACTIN7 was used as an internal standard to normalize the amplification products. The following primers were used: FLC, forward 5′-CTGTTCTCTGTGACGCATCC-3′ and reverse 5′-AGTGTGAACCATAGTTCAGAGC-3′; AtCYP707A1, forward 5′-CCATCGCTCAAGACTCTCTC-3′ and reverse 5′-CCTCGTCTTTTCCGAAGATCG-3′; AtCYP707A2, forward 5′-CAATTCCTTCTTCGCCACTCG-3′ and reverse 5′-GCCTCTGGTCCAATCATACG-3′; AtNCED9, forward 5′-CGACCGGAGAGATTCGAAAG-3′ and reverse 5′-CACCTTCTCCTCGTCGTGAAC-3′; AtABI3, forward 5′-ACGTCAGCAGGTGGTACCAG-3′ and reverse 5′-GGCAAGTGTGTCTCAGCTTC-3′; AtABI4, forward 5′-ATGGACCCTTTAGCTTCCCAAC-3′ and reverse 5′-CTTTGCGTTTGCGTTGAGCG-3′; AtABI5, forward 5′-AGTCTGCTGCTAGATCTAGAG-3′ and reverse 5′-TGTTGCTTCCTCTTCCTCTCC-3′; AtGA3ox1, forward 5’-TTCACCATCACTGGCTCGCC-3’ and reverse 5′-CCACATCAATTTCGATGCCAAC-3′; AtGA3ox2, forward 5′-TAAGAAGATGTGGTCCGAAGG-3′ and reverse 5′-CACATCAACTTGGCTGCCAAC-3′; AtACTIN7, forward 5′-AGGGAGAAGATGACTCAGATC-3′ and reverse 5′-GTGTGAGACACA CCATCACC-3′.

Construction of pFY::GUS transgene and GUS histochemical staining

The promoter of AtFY was amplified by PCR with genomic DNA from Arabidopsis (Columbia ecotype), using the primers PAtFY-F, 5′-AAGCTTCTCGGTGTGTTTTTGGGG-3′ and PAtFY-R, 5′-GGATCCCATTGTTGCCCACGGAG-3′. The 1.33 kb PCR product was amplified and cloned into pGEM-Teasy vector (Promega,http://www.promega.com) and confirmed by sequencing. The FY promoter fragment was then introduced upstream of β-glucuronidase (GUS) at the HindIII/BamH1 site by replacing the CaMV 35S promoter in the pBI121 binary vector (Clontech, http://www.clontech.com). The pFY::GUS was mobilized into Agrobacterium GV3101 and then introduced into Arabidopsis by the floral dip transformation (Clough and Bent, 1998). Transformed lines were selected using kanamycin (50 μg ml−1) in MS medium with 1% sucrose. Plants were stained for GUS expression as described (Stomp, 1992). Briefly, the tissues were incubated overnight at 37 °C in staining buffer (100 mM phosphate buffer, pH 7.5, 10 mM EDTA, 0.1% TritonX-100, 0.5 mM K4Fe(CN)6, 0.5 mM K3Fe(CN)6, and 0.5 mg ml−1 X-gluc). Seeds were dissected into the embryo and endosperm under a dissecting microscope before staining. After staining, tissues were transferred to a solution of ethanol: acetic acid (3:1 v/v) and photographed using a Leica DC500 dissecting microscope (http://www.leica-microsystems.com).

Construction of 35S::FY/AAFY complementation transgene

The opening reading frame of FY was amplified using the cDNA templates prepared from Ler leaves with the primers Nflag-F (5′-TCTAGAATGG ACTACAAAGACGATGACG-3′) and FY3Xb-R (5′-TCTAGACTACTGATGTTGCTGATTGTTG-3′). The two PPLPP domains were mutagenized to AAAPP by using two steps of overlapping PCR with the following primers: Nflag-F (5′-TCTAGAATGGACTACAAAGACGATGACG-3′); 1PP2AA-F (5′-GGGCTGCAGCTGCACCACCTGGTCCCCACCCATCG-3′); 1PP2AA-R (5′-GTGGTGCAGCTGCAGCCCCCAGTGCCATGGAACCTGGAAG-3′); 2PP2AA-F (5′-AGCAGGCAGCTGCACCTCCTGGCCCTCCACCAAAC-3′); 2PP2AA-R (5′-GAGGTGCAGCTGCCTGCTGCTGCTGTTGGAAAG-3′); and FY3Xb-R (5′-TCTAGACTACTGATGTTGCTGATTGTTG-3′). The mutant FY with two AALPP instead of two PPLPP was designated as AAFY. The PCR products were cloned into pGEM-Teasy (Promega, http://www.promega.com) and sequenced. The target genes (FY/AAFY) were subcloned into pFGL 779, a derivative of pPZP201 at the XbaI site. The 35S::FY/AAFY were transformed into the Arabidopsis fy-1 mutant by Agrobacterium-mediated floral dip transformation (Clough and Bent, 1998). Glufosinate-ammonium (10 mg l−1) was used for selecting transformed lines.

Analysis of ABA

ABA levels were measured using the Phytodetek competitive ELISA kit (Agdia, Elkhardt, IN, USA) following the manufacturer’s instructions. Tissue was powdered and extracted overnight in 1 ml of ABA extraction buffer (methanol, containing 100 mg l−1 butylated hydroxyl toluene, 0.5 g l−1 citric acid monohydrate). Samples were then centrifuged at 2500 g and the supernatant was dried with N2. Dry extracts were re-suspended in 100 μl pure methanol followed by the addition of 900 μl of 1× TRIS-TBS buffer. For ABA ELISA 10–20 μl of the final extracts were further diluted in 1× TBS in order to obtain a final volume of 100 μl. Analysis was then conducted following the Phytodetek protocol. ABA levels were expressed as pmol mg−1 dry tissue.

Results

Arabidopsis fy-1 displayed a lack of seed dormancy and decreased sensitivity to ABA during germination

Freshly-harvested seed of fy-1 and Ler were imbibed on water-soaked filter paper, devoid of nutrients, under fluorescent light in the growth chamber for a period of 5 d. At the end of 5 d, seeds of the fy-1 mutant germinated while most of the Ler seed failed to germinate under the same conditions (Fig. 1A). The fy-1 seed completed germination by the third day of imbibition, while Ler showed approximately 10% germination after 5 d. Ler seeds required 4–5 weeks of dry after-ripening to reach complete germination after 5 d imbibition, while fy-1 seeds germinated readily without after-ripening (Fig. 1B).

Fig. 1.

Germination characteristics of Ler (wild type) and fy-1. (A) Germination (%) of freshly-harvested Ler and fy-1 mutant seeds. (B) Germination (%) of Ler and fy-1 mutant after different periods of dry storage. Seed germination was scored after 5 d imbibition. (C) Germination (%) of stratified Ler and fy-1 seeds at 5 d in the presence of ABA. (D) Changes in radicle length of isolated embryos of Ler and fy-1 in the presence of 4 μM ABA. The imbibed seeds were kept at 4 °C for 3 d before embryo isolation. (E) Stratified fy-1 seed at 7 d imbibition with 3 μM ABA (a) or H2O (b). Germination frequencies are means of triplicate assays. Radicle lengths are means of 15 embryos for each genotype. Error bars represent ±1.96 SE. (This figure is available in colour at JXB online.)

Stratified seed of fy-1 was less sensitive to ABA than stratified Ler (Fig. 1C). Ninety per cent inhibition of Ler germination was attained with 3 μM ABA, while fy-1 germination was inhibited by 25%. Embryo radicle elongation also showed a differential response to ABA between Ler and fy-1 (Fig. 1D). Post-germination growth of the fy-1 mutant was, however, affected by 3 μM ABA (Fig. 1Ea), as the untreated seedlings continued to develop producing green cotyledons (Fig. 1Eb).

PPLP repeats within the carboxyl terminus of FY are not required for acquisition of seed dormancy and ABA sensitivity

Complementation studies were performed, introducing the FY open reading frame of ∼2.2 kb, driven by the constitutive CaMV 35S promoter, into an fy-1 loss-of-function background. Two independently transformed lines (C2 and C6) were selected on glufosinate–ammonium in the T2 generation showing a 3:1 segregation ratio indicating a single insertion of 35S::FY transgene in the genome. The homozygous lines (C2, C6), obtained in the T3 generation, were used for further functional analysis. Northern blot analysis from the total RNA obtained from freshly harvested dry seeds revealed a higher level of expression of FY in the two transgenic complemented lines (C2 and C6) when compared to Ler and fy-1 mutant (Fig. 2A). The fy-1 mutant is known as a late-flowering phenotype (Koornneef ; Simpson ) and total leaf number at flowering is correlated with flowering time (Koornneef ). Both complemented lines possessed similar leaf numbers as Ler at flowering (Fig. 2B). Thus, the introduced FY gene restored the flowering phenotype and was fully functional.

Fig. 2.

Effect of introduction of FY into fy-1 on seed dormancy. (A) Transcript level of FY in Ler, fy-1, and complementation lines (C2, C6). RNA was extracted from freshly-harvested seed. Each lane contained 10 μg total RNA. (B) Total leaf number of Ler, fy-1, and complementation lines (C2, C6) at the time of flowering. Values are means of 10 plants and error bars represent ±1.96 SE. (C) Germination (%) of freshly-harvested seeds from Ler, fy-1, and complementation lines (C2, C6). (D) Germination (%) of Ler, fy-1, and complementation lines (C2, C6) with 3 μM ABA. The imbibed seeds were kept at 4 °C for 3 d before moving to a growth chamber. Germination was scored at 5 d. Germination percentages are means of triplicate and error bars represent ±1.96 SE.

Using freshly harvested seeds, the complemented lines showed a similar level of germination (∼10%) to Ler (Fig. 2C) after 5 d of imbibition, while the fy-1 mutant germinated completely. Also, both complemented lines (C2, C6) restored ABA sensitivity of fy-1 with a similar germination percentage to Ler (∼10%) when imbibed with 3 μM ABA after 5 d of imbibitions (Fig. 2D). These results demonstrate FY involvement in the acquisition of dormancy and ABA sensitivity in Arabidopsis seeds.

Arabidopsis FY contains seven conserved WD repeats at the N-terminus and two non-conserved domains (PPLPP) at the C-terminal region. It has been reported that these two PPLPP domains are required for in vitro FY–FCA interaction and floral transition (Henderson ). The fy-1 mutant lacks these PPLPP domains due to the shift in the splice acceptor site within exon 16, resulting in a premature stop codon (Simpson ). In order to determine whether these two PPLPP domains of FY are important to acquire ABA sensitivity and dormancy, the first two proline (P) residues of both PPLPP domains were each mutated to alanine (A) and a construct, in which both PPLPP domains were mutated, was generated. The mutant, AAFY, driven constitutively by the CaMV 35S promoter, was introduced into fy-1. Two homozygous transgenic lines (AAC1 and AAC3) in the T3 generation were obtained, based on glufosinate–ammonium selection. Northern blot analysis detected the presence of AAFY transcripts (Fig. 3A). Seed germination was assayed using freshly-harvested seeds from AAC1 and AAC3 transgenic plants along with Ler and the fy-1 mutant. Both seed dormancy and ABA-insensitivity of the fy-1 mutant were complemented by both transgenic lines expressing AAFY (Fig. 3B, C). However, neither of the transgenic lines expressing AAFY could rescue the late-flowering phenotype of the fy-1 mutant (Fig. 3D). This suggests that intact PPLPP repeats are not required for FY function in seed dormancy and ABA sensitivity, but are essential in floral transition.

Fig. 3.

PPLPP repeats of FY are not important for seed dormancy. (A) Northern blot showing the expression of FY in Ler, fy-1, and homozygote transgenic lines with 35S::AAFY (AAC1, AAC3). RNA was extracted from freshly-harvested dry seeds. Each lane contained 10 μg total RNA. (AAFY: FY with 2× PPLPP→AALPP). (B) Germination (%) of fresh seeds from Ler, fy-1, AAC1, and AAC3. The seeds were scored after 5 d imbibition on filter paper. Germination percentages are means of triplicate and error bars represent ±1.96 SE. (C) Germination (%) of Ler, fy-1, AAC1, and AAC3 in the presence of 3 μM ABA. The imbibed seeds were kept at 4 °C for 3 d before moving to a growth chamber. Germination was scored at 5 d. Germination percentages are means of triplicate and error bars represent ±1.96 SE. (D) Total leaf number at flowering of Ler, fy-1, AAC1, and AAC3. Values are means of 10 plants and error bars represent ±1.96 SE.

The interaction of FCA–FY affects flowering by down-regulating FLC levels in leaves (Simpson ) and it has recently been shown that FLC affects temperature-dependent germination in Arabidopsis (Chiang ). Real-time RT-PCR of total RNA extracted from freshly-harvested dry seeds of Arabidopsis genotypes Ler, fy-1, C2, C6, AAC1, and AAC3 showed that the AAFY transgenic lines (AAC1and AAC3) had similar levels of FLC to the fy-1 mutant while FLC levels were reduced in FY complementation lines (C2 and C6) and Ler (Fig. 4). A similar trend was observed in Northern analysis (see Supplementary Fig. S1 at JXB online).

Fig. 4.

FLC expression in Arabidopsis genotypes. Real-time RT-PCR of FLC in freshly-harvested seed of Arabidopsis genotypes Ler, fy-1, C2, C6, AAC1, and AAC3. The phenotypes associated with the lines are indicated.

Effect of FY on genes involved in ABA signalling and GA metabolism

The regulation of seed germination is determined by a balance between the phytohormones, ABA and GA (Debeaujon and Koornneef, 2000; Finkelstein ). Genes involved in ABA and GA signalling and metabolism were analysed to determine whether there was altered expression of these genes in the fy-1 mutant and how ABA influenced this expression. Three time periods of up to 48 h imbibition time, just prior to the first appearance of germinated seed, were chosen. The genes CYP707A1, CYP707A2, NCED9, GA3OX1, GA3OX2, ABI3, ABI4, ABI5, and RGL2 were selected for testing. CYP707A1 and CYP707A2 encode an ABA 8'-hydroxylase, responsible for ABA catabolism (Okamoto ) and NCED9 is required for ABA biosynthesis in seeds (Lefebvre ). GA3ox1 and GA3ox2 are two central GA biosynthesis genes that form bioactive GA in Arabidopsis (Yamaguchi ; Yamauchi ). RGL2 codes for a DELLA protein that is a GA signal repressor inhibiting germination by stimulating ABA and ABI5 synthesis (Piskurewicz ). ABI3, ABI4, and ABI5 are ABA signalling genes that are involved in the regulation of seed germination by ABA (Finkelstein and Lynch, 2000; Söderman ). ABA biosynthesis and catabolism genes were not significantly different between Ler and the fy-1 mutant over the 48 h of imbibition (Fig. 5). ABI3 was not significantly different over the same period, but ABI4 was significantly higher in fy-1 at 48 h imbibition. ABI5 was consistently lower in fy-1 over the course of the imbibition and significantly lower after 1 d of imbibition (Fig. 5). GA3ox1 increased throughout imbibition in fy-1 and was significantly higher in the fy-1 mutant from 24 h imbibition (Fig. 5). GA3ox2 began to increase rapidly in fy-1 after 24 h imbibition and was significantly higher than Ler by 48 h imbibition (Fig. 5). There were no changes in expression of either gene in Ler seeds throughout imbibition. The differences between Ler and fy-1 in the GA3ox genes are characteristic of the differences between dormant and non-dormant seeds (Finch-Savage and Leubner-Metzger, 2006), consistent with an effect of FY on dormancy. RGL2 levels increased during imbibition of freshly-harvested seeds of both Ler and fy-1, with fy-1 having significantly higher transcript levels than Ler at 24 h and 48 h imbibition.

Fig. 5.

Expression of ABA- and GA-related genes during imbibition in Ler and fy-1. Transcript levels of CYP707A1, CYP707A2, NCED9, ABI3, ABI4, ABI5, GA3ox1, GA3ox2, and RGL2 analysed by quantitative RT-PCR. CDNA was prepared from freshly-harvested seeds of Ler and fy-1 during the first 48 h imbibition. The expression value of the individual gene was normalized to ACTIN7 as an internal control. Values, with the exception of ABI5 are means from three replicates from different seed lots. The data for ABI5 are means from four different seed lots. Error bars represent ±1.96 SE.

ABA levels were measured in freshly-harvested seed to determine whether changes in ABA could explain the differences in dormancy between Ler and fy-1 and the changes observed in the examined transcripts during imbibition. ABA levels were significantly higher in fy-1 than Ler at the start of imbibition and remained higher, although with considerable variability between replicates, in fy-1 at all sample times (Fig. 6). Ler ABA moderately declined throughout imbibition while a sharp decline in fy-1 ABA levels occurred in the interval between 24 h and 48 h imbibition.

Fig. 6.

ABA levels in freshly-harvested, imbibed Ler and fy-1. Freshly-harvested seed of Ler and fy-1 were imbibed in water for various time periods and their ABA content measured. Three biological replicates of each line and time point were analysed. The results are expressed as means ±1.96 SE.

Expression of ABI5 in relation to FY during seed germination

In view of the differences in ABI5 transcripts in the two lines, ABI5 transcripts were examined in freshly-harvested seed of Ler, fy-1, and the previously described complemented lines, C2, C6, AAC1, and AAC3, that had rescued seed dormancy (Fig. 4). If the restoration of seed dormancy in the complemented lines was the result of restored ABI5, this should be apparent from the levels of ABI5 in the tissue. Complementation of fy-1 with either complete FY or PPLPP-mutated FY restored ABI5 levels to that of the wild type, Ler (Fig. 7).

Fig. 7.

ABI5 levels in FY-complemented fy-1. Freshly-harvested seed of Ler, fy-1, and the FY-complemented lines C2, C6, AAC1, and AAC3 were imbibed in water for 24 h prior to extraction of RNA and real-time RT-PCR. Three replicates of each line were analysed and the results are presented as means ±1.96 SE.

Stratified Ler and fy-1 seeds were examined to determine how the two non-dormant lines responded, relative to FY and ABI5 expression, to an ABA concentration sufficient to repress their germination. Stratified seed imbibed in water germinated rapidly, with complete germination occurring within 48 h. Ten μM ABA completely inhibited this germination. After stratification, Fy levels in Ler were about 6-fold higher than in fy-1 (Fig. 8A). Over the 72 h germination at 22 °C, there was no change in the Fy levels of the fy-1 mutant, irrespective of the presence of ABA. In the case of Ler stratified seeds imbibed in water, Fy levels declined to the levels in fy-1 seed over the first 48 h of imbibition, remaining constant over the remaining 24 h. For Ler seeds imbibed in ABA, FY levels were not significantly different from water-imbibed seed over the first 24 h, but were higher than water-imbibed seed over the remaining 48 h of measurement. ABI5 levels in stratified seeds (Fig. 8A) were 5–20% of those in non-stratified seeds (Fig. 5). The ABI5 levels in ABA-treated seeds were 2.5-fold higher than those in water-treated seeds at the end of stratification (Fig. 8A). The ABI5 levels were relatively constant and similar for Ler and fy-1 over the course of the imbibition for water-treated seed. For ABA-treated seed, the two lines had the same ABI5 levels at the start of the incubation. During the first 24 h in the presence of ABA, Ler ABI5 levels remained constant, while fy-1 ABI5 levels declined. After 24 h in ABA, the two lines showed parallel increases in ABI5 levels over the remainder of the incubation, with the fy-1 levels being consistently lower.

Fig. 8.

FY and ABI5 expression during seed germination. FY and ABI5 transcript levels in stratified Ler and fy-1 seeds during imbibition in H2O or 10 μM ABA. The values are means ±standard deviation from two biological replicates. (B) Tissue expression of FY::GUS in 1 d imbibed seeds.

Histochemical analysis, using an FY native promoter fused to GUS (pFY::GUS) showed intense GUS staining in both the embryo and endosperm after 24 h imbibition of transgenic seeds in water (Fig. 8B).

Discussion

A component of the autonomous flowering pathway, FY, along with its interacting partners FCA, is known to regulate flowering in Arabidopsis (Simpson ). FY was identified through its interaction with FCA, and its molecular function was thought to be dependent on its interaction with FCA (Quesada ). Studies with fy mutants suggested that FY can affect various physiological functions (Henderson ) through the dynamic interaction complexes it can form with or without FCA (Quesada ; Manzano ).

The work reported here demonstrates that FY expression also affects seed dormancy and ABA sensitivity to germination in Arabidopsis seeds and that the mechanism by which FY accomplishes this affect differs from the mechanism by which FY regulates flowering in the autonomous pathway.

ABA insensitivity in Arabidopsis fy-1seeds

Freshly-harvested fy-1 seeds germinated rapidly (Fig. 1A), whereas Ler required 5 weeks of dry storage to achieve the same germination capacity as freshly-harvested fy-1 (Fig. 1B). The two hormones, GA and ABA, are critical regulators of germination (Bentsink and Koornneef, 2008), with ABA considered to impose a block on germination. The differences in germination do not appear to be the result of ABA repression of germination as the ABA levels of freshly-harvested fy-1 were significantly higher, not lower, than Ler ABA at the start of imbibition (Fig. 6) and tended to be higher throughout the imbibition period. ABA levels in fy-1 were considerably more variable than Ler, probably reflecting variability in the effect of environment on seed status of the biological replicates relative to germination and dormancy in the non-dormant fy-1. ABA insensitivity in the fy-1 mutant is the more likely cause of the phenotype as stratified fy-1 seed had significantly less sensitivity to ABA than stratified Ler (Fig. 1C) and fy-1 embryo radicle growth was significantly less inhibited by ABA than Ler (Fig. 1D). The germination phenotype of the wild-type parent can be rescued in both first generation progeny from reciprocal crosses between the mutant and Ler (see Supplementary Fig. S2 at JXB online). This indicates nuclear control and suggests that ABA insensitivity in the mutant might be found in both the endosperm and embryo of the seed. FY is expressed both within the embryo and the endosperm of 1 d imbibed seeds (Fig. 7B). During seed development, Arabidopsis FY expression is confined to the embryo at the heart stage of development but is evident in both embryo and endosperm in the mature seed (Henderson ). The effect of FY expression on seed ABA sensitivity may, therefore, be the result of an event affected by FY during late embryogenesis and/or during stratification.

ABA catabolism and GA synthesis are considered key elements in controlling germination and dormancy (Debeaujon and Koornneef, 2000; Millar ). There were no significant differences in expression of two enzymes of ABA catabolism, CYP707A1 and CYP707A2, between freshly-harvested, imbibed seed of Ler and fy-1 (Fig. 5). Similarly, there were no significant differences between Ler and fy-1 during imbibition for NCED9, an enzyme involved in ABA synthesis (Lefebvre ) in Arabidopsis seed.

Low GA levels result in an increase in ABA through stabilization of RGL2, a DELLA factor that represses germination. (Piskurewicz ). What is the relationship between FY, RGL2, and the regulation of germination? RGL2 is low at the start of imbibition, with transcript levels increasing in both lines over the course of imbibition (Fig. 5), as increased metabolic activity commences. FY, as a 3'-polyadenylation factor, stabilizes transcript stability (Zhao ) and if RGL2 stability were a factor in the ABA insensitivity of fy-1, then RGL2 levels would be predicted to be lower than Ler over the course of the imbibition and they are not. The higher relative transcript levels of the gene in fy-1 during imbibition likely reflect the tendency towards higher levels of ABA in fy-1 (Fig. 6), since ABA affects RGL2 expression (Piskurewicz ). The significant increases in GA3Ox1 and GA30x2 in fy-1 compared with Ler in the first 24 h of imbibition (Fig. 5) suggest that significant increases in fy-1 GA levels might be expected. This should result in increased protein turnover of RGL2 (Piskurewicz ) in fy-1, but not in Ler, during imbibition, which should relieve the RGL2 repression which is observed (Fig. 1A). It has been suggested, however, that ABI5 acts as the final common repressor of germination in response to changes in ABA and GA levels (Piskurewicz ) and that ABI5 levels are higher in Ler at 24 h and 48 h of imbibition (Fig. 5). In this scenario, ABI5 would be the controlling factor regulating germination status. It is evident (Fig. 6) that a decline in ABA is not essential for the initiation of germination processes in fy-1.

Of the three ABA response transcription factors tested, ABI3, ABI4, and ABI5, only ABI5 had consistently lower transcript levels during imbibition, with the differences being significant at the 24 h and 48 h imbibition periods (Fig. 5). There are two possible considerations as to why there is no significant difference in ABI5 levels at the start of imbibition. The first possibility is that changes in ABI5 that affect dormancy occur during imbibition when differences in the metabolism (Millar ) or amount (Ali-Rachedi ) of ABA between dormant and non-dormant seed influence ABA-induced synthesis of ABI5 (Finkelstein and Lynch, 2000; Lopez-Molina and Chua, 2000). Our examination of a few of the enzymes involved in ABA catabolism and biosynthesis (Fig. 5) are not definitive enough to support this possibility. The second possibility is that there is a relationship between ABI5 levels in freshly harvested seed and seed dormancy, but the variable growth and harvest conditions of the various seed lots mask this effect in the biological replicates. Regardless of the lack of difference in the ABI5 levels at time zero, the levels at 24 h and 48 h imbibition are significantly higher in Ler (Fig. 5) and it is essentially dormant, while fy-1 is in the process of germination (Fig. 1A). There are 5–15-fold differences in ABI5 levels between dormant (Fig. 5) and non-dormant seed (Fig. 8), with the stratified seed having much lower levels. ABA promotes ABI5 expression (Fig. 8), in agreement with previous work (Finkelstein and Lynch, 2000; Lopez-Molina and Chua, 2000). This is also consistent with the tendency of ABI5 to be lower in freshly-harvested fy-1, even though fy-1 ABA levels tend to be higher than those in Ler (Fig. 6), an effect that could be the result of lack of ABI5 transcript stability in fy-1. ABI5 is expressed in the embryo and the micropylar region of the endosperm (Penfield ) and has been suggested to be the final common repressor of germination in response to changes in ABA and GA (Piskurewicz ), acting downstream of ABI3 (Lopez-Molina ). Furthermore, complementation of fy-1, with either fully-functional FY or PPLPP-mutated FY rescued both seed dormancy and wild-type ABI5 levels.

ABA increased ABI5 levels, as anticipated from previous studies (Lopez-Molina ; Finkelstein and Lynch, 2000), with no substantial changes in FY levels in stratified seeds imbibed in the absence and presence of ABA at sufficient concentrations to repress germination in both lines (Fig. 8). This is counterintuitive to what might be expected if ABI5 levels were totally dependent on FY for transcript stability. ABI5 may have been maintained as the result of ABA-induced increased ABI5 transcription to offset the increased breakdown due to the inadequate 3'-polyadenylation. Furthermore, germination repression may be the result of the combined effects of ABA and ABI5 as it has been shown that the two can act synergistically in ABA-regulated events (Gampala ).

Based on the above considerations, ABI5 transcript stability is a likely factor resulting in decreased ABA sensitivity of fy-1.

FY in relation to seed germination and flowering in Arabidopsis

Acknowledging that fy-1seeds do have an ABA insensitivity, possibly as a result of modified ABI5 expression, does the mutation in FY that results in altered flowering in Arabidopsis operate via a similar mechanism to produce seed ABA insensitivity? Although there remain many unanswered questions, there is enough information to conclude that the pathways are different.

Complementation of fy-1 with wild-type FY rescued both the germination and flowering phenotype (Fig. 2). PPLPP repeats in the C-terminal region of FY are important for regulating FLC levels in vegetative rosette tissue (Simpson ). The introduction of PPLPP-mutated FY (AAFY) into a fy-1 loss-of-function background rescued inhibition of seed germination and ABA sensitivity, but not flowering in fy-1 (Fig. 3), suggesting FY functions differently in seed germination than in the transition to flowering in the autonomous pathway. This is supported by the evidence that complementation of fy-1 with wild type FY reduced FLC, late flowering, and inhibition of seed germination, whereas complementation with PPLPP-mutated FY (AAFY) in the transgenic fy-1 mutant line had little effect on FLC or late flowering but still affected seed dormancy (Fig. 4). These findings suggest that the effect of FY on seed germination is not processed via FLC expression and that the PPLPP domains of FY are not required to regulate seed germination through ABA. FY interacts with FCA via the PPLPP motifs in the C-terminal region of FY and the WW domain of FCA, and this interaction is important for floral transition (Simpson ). FY, but not FCA, has a pleiotropic function in embryo development (Henderson ). The FCA–FY interaction is transient in Arabidopsis (Manzano ). PPLPP-mutated FY (AAFY), expressed in fy-1, is unable to interact with FCA, yet the expression of the mutated gene in fy-1 rescued ABA-regulated seed germination, indicating that FY function relative to seed germination may not depend on FCA.