Introduction

Flowering is an important process in the transition from vegetative to reproductive growth in plant, and is regulated by intricate networks of both endogenous and environmental cues (Simpson and Dean 2002; Baurle and Dean 2006). In Arabidopsis, the identification and characterization of flowering-defective mutants have led to the identification of at least five signaling pathways that co-regulate flowering, such as photoperiod, autonomous floral initiation, vernalization, brassinosteroid (BR) and gibberellic acid (GA) pathways (Boss et al. 2004; Simpson 2004; Amasino 2005; Clouse 2008; Michaels 2009; Li et al. 2010).

Mutations in CONSTANS (CO), GIGANTEA (GI) and FLOWERING LOCUS T (FT) cause late-flowering phenotype which are insensitive to photoperiod (Putterill et al. 1995; Koornneef et al. 1998; Kardailsky et al. 1999). FLOWERING LOCUS KH DOMAIN (FLK), LUMINIDEPENDENS (LD), FVE and FLOWERING LOCUS D (FLD) are involved in the autonomous pathway regulating flowering (Lee et al. 1994; He et al. 2003; Ausin et al. 2004; Lim et al. 2004). FRIGIDA (FRI) and VERNALIZATION INSENSITIVEs (VANs) play major roles in conferring a requirement for vernalization to initiate flowering (Johanson et al. 2000; Gendall et al. 2001; Levy 2002; Sung and Amasino 2004). GA signals play a major role in promoting flowering especially under short day (SD) conditions in Arabidopsis. Mutants blocked in GA biosynthesis and defective in GA signaling, such as gibberellin deficient 1 (ga1), gibberellin insensitive (gai), and spindly (spy), were found to be delayed in flowering (Wilson et al. 1992; Jacobsen and Olszewski 1993; Wilson and Somerville 1995). Recently, several studies demonstrated that these flowering pathways mainly converge on FLOWERING LOCUS C (FLC) or SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1 (SOC1), which are two MADS box transcriptional factors (Michaels and Amasino 1999; Lee et al. 2000; Li et al. 2016). FLC acts as a strong repressor of initiation of flowering upstream of FLOWERING LOCUS T (FT) and SOC1 (Searle et al. 2006). FRI is a positive regulator that increases the FLC expression level (Johanson et al. 2000). In addition, many genes involved in vernalization and autonomous pathways can control the epigenetic status of FLC chromatin and reduce FLC expression, resulting in early flowering (Amasino 2004; Baurle and Dean 2006). A recent report has shown that DELLA proteins interact with FLC to enhance its inhibition ability. GA can promote the degradation of DELLA to reduce the inhibition ability of FLC, leading to early flowering (Li et al. 2016). GA can also positively regulate SOC1 expression under SD conditions (Moon et al. 2003).

BRs constitute a class of steroid hormones that regulate various aspects of plant development including photomorphogenesis, xylem formation, cell division and elongation (Li 2010). Several studies revealed that BR-deficient and insensitive mutants det2, dwf4, and bri1, showed a late-flowering phenotype, and the bas1 sob7 double mutant in metabolizing BRs exhibited an early flowering phenotype (Chory et al. 1991; Azpiroz et al. 1998; Turk et al. 2005; Domagalska et al. 2007). The late-flowering phenotype is accompanied by an increase in FLC expression (Domagalska et al. 2007). BR signals result in a chromatin modification which requires ELF6 and REF6 (Yu et al. 2008). BR signaling also interacts with GA and abscisic acid (ABA) pathways signals to control flowering time (Domagalska et al. 2010).

In addition, ABA, ethylene, ambient temperature, and light quality are also critical in regulating flowering (Gray 2004; Achard et al. 2007; Lee et al. 2007; Wollenberg et al. 2008; Davis 2009; Kotchoni et al. 2009). Although the major molecular players and pathways regulating floral transition have been identified, further components remain to be discovered.

Sterols play significant roles as components of the cell membrane and as biosynthetic precursors of steroidal hormones in both animals and plants (Clouse 2000; Gilbert et al. 2002; Attard et al. 2009). In a previous report, rapid changes in sterol and phospholipid/sphingolipid composition led to high activity in the floral apex. These processes might play an important role in regulating the initiation of flowering (Hobbs et al. 1996). In Arabidopsis, campesterol, which is one species of the final sterols, is the precursor of BRs (Clouse 2000). A number of proteins and metabolites of post-squalene sterol biosynthesis (SQE1, SMT1, CPI1, CYP51A2, FK, HYD1, and SMT2) play additional roles in plant development independent of that of sterols as the precursors of brassinosteroids (Schrick et al. 2000, 2002; Carland et al. 2002; Souter et al. 2002; Kim et al. 2005; Men et al. 2008; Pose et al. 2009). FACKEL (FK) is a C-14 reductase involved in the early steps of sterol biosynthesis. The previously reported FK mutant lines, such as fk-X224 and fk-J79, exhibit sterile and/or lethal phenotypes because of severe defects in embryogenesis and post-embryonic development (Jang et al. 2000; Schrick et al. 2000). More recently, Qian et al. (2013) identified a weak mutant allele of FK that is defective in sterol biosynthesis, fk-J3158, which displays uncontrolled cell fate commitment and maintenance of the stomatal lineage in Arabidopsis. Another FK mutant allele fk-J79, which is also known as extra-long-lifespan1 (ell1), was originally characterized for its long lifespan. Likewise, fk-J3158 also displays a late-flowering phenotype. However, the specific mechanisms are still unclear. Here, we provide physiological and genetic data on the weak allele fk-J3158 to demonstrate that FK acts as an additional pathway involved in GA signaling and vernalization for flowering in Arabidopsis.

Materials and methods

Plant materials and growth conditions

The Arabidopsis thaliana Columbia-0 (Col-0; seeds obtained from Arabidopsis Biological Resource Center, ABRC, Columbus, OH, USA) was used as a wild-type control. Mutants and transgenic plants used in this study are as follows: fk-J3158, ProFK::FK/fk-J3158, 35S::FK, ga1-3(CS3104), flc(SALK_140021), fk-J3158 flc and fk-J3158 ga1-3. The mutation locus in fk-J3158, ga1-3 and flc background was sequenced (for primer sequences, see Table S1). Seedlings were germinated on 1/2 MS agar plates (Murashige and Skoog 1962) and transferred to soil for growth with 16 h light/8 h dark cycles (LD conditions) or 8 h light/16 h dark cycles (SD conditions) in 22 °C. ga1-3(CS3104) and SALK_140021 was obtained from Arabidopsis Biological Resource Center (ABRC, Columbus, OH, USA).

Imaging and measurement

Images of plants were documented using a Nikon COOLPIX 7000 camera. Flowering time was scored as the number of rosette leaves at flowering when the bolt was approximately 1 cm high. A total of 20–60 plants per genotype were analyzed in each experiment. For statistical analysis, values are reported as the mean ± standard deviation, and unpaired Student’s t test was performed.

GA3 and paclobutrazol (PAC) treatment

To the plants in soil, the plants were sprayed by a 100 µM GA3 (Sigma–Aldrich, St. Louis, MO, USA) or 37 mg/l PAC (Sigma) solution with 0.02% Tween20 under LD conditions until bolting (twice in 1 week). To the plants in MS plate, the media was contained different concentrations of GA3, 1, 10, 50, and 100 µM. To real-time RT-PCR, the seedlings from MS plate were put into 100 µM GA3 or 10 µM PAC solutions shaken for 2 h in dark.

RNA extraction and quantitative real-time polymerase chain reaction (qRT-PCR)

Total RNA was isolated from the 3rd or 4th rosette leaves of 15-day seedlings using the Plant RNA Kit (Omega Bio-tek, Norcross, GA, USA) according to the manufacturer’s instruction. To analyze FLC and SOC1 gene expression, total RNA was isolated at the 16th hour of 16 h light in LD conditions and 8th hour of 8 h light in SD conditions. One microgram of total RNA was used for the reverse transcription using a PrimeScript RT reagent Kit (Takara Bio Inc., Kusatsu, Shiga, Japan) according to the manufacturer’s protocol. qRT-PCR was performed using SYBR Premix EX Taq II (Takara) on a Stratagene Mx3000P PCR instrument (for primer sequences, see Suppl. Table S1). The glyceraldehyde-3-phosphate dehydrogenase C subunit (GAPC) gene was used as a control gene for cDNA amount in the real-time PCR.

GUS reporter gene histochemical analysis

The GUS histochemical analysis of reporter gene was performed on generation seedlings of two independent T1 lines. 24 h staining for GUS activity was performed on samples representing various stages of plant development including 3-day-old whole seedling, 25-day-old whole plant, maturing buds and flowers using a GUS staining solution (Jefferson et al. 1987).

Vernalization treatment

Seeds were planted on MS plates for 3 weeks in 4 °C and in dark. These seedlings were transferred to 16 h light/8 h dark cycles (LD conditions) in 22 °C until the leaves became green. Then they were transferred to soil.

Results

The fk-J3158 mutant in early sterol biosynthesis delays the flowering time

In a previous study, we reported that the mutations of FK affected early sterol biosynthesis and resulted in defects in stomatal development and patterning (Qian et al. 2013). Among all the currently available FK mutants, fk-J3158 is the only weak mutant allele because of its fertility in soil and its weak mutant phenotypes in both stomatal development and other organs as compared to the other FK mutants (Qian et al. 2013). Based on the multiple defects of fk-J3158, it was suggested that FK functions in several processes of plant growth and development. For this study, we explored the dwarfed growth and long lifespan phenotype of fk-J3158. The defective development in fk-J3158 mutant was complemented by the expression of FK cDNA driven by the FK own promoter (Fig. 1a–c). To further evaluate the function of FK in the regulation of flowering, a 35S::FK line was created whereby FK was over-expressed in Col (Fig. 1d, e). Under LD conditions, the average bolting days of wild type and fk-J3158 plants were 33.29 and 48.73, and the average total leaf number were 9.72 and 15.25, respectively (Fig. 2a, c, d). The average bolting days of ProFK::FK/fk-J3158 were similar to Col and earlier than fk-J3158 (Fig. 2a, c). The average total leaf number of ProFK::FK/fk-J3158 was significantly less than fk-J3158 and slightly more than Col (Fig. 2a, d). The 35S::FK plants bolting with fewer days and leaves (30.57 and 9.82, respectively) compared to wild type under the LD photoperiod conditions, whereas it displayed normal plant growth and development (Fig. 2a, c, d). In order to understand whether FK functions in the photoperiod pathway, the Col, fk-J3158, ProFK::FK/fk-J3158 and 35S::FK plants were also grown under SD conditions. We found that 35S::FK, ProFK::FK/fk-J3158, and fk-J3158 plants all exhibited significantly delayed time to bolting and increased total leaf number similar to Col plants (Fig. 2b, e, f). Thus, these results suggested that FK functions to promote flowering and regulates flowering time independent of the photoperiod pathway.

Fig. 1
figure 1

Seedling phenotypes of Col-0, fk-J3158, ProFK::FK/fk-J3158 and 35S::FK. Seedling phenotypes in 15-day-old Col-0 (a), fk-J3158 (b), ProFK::FK/fk-J3158 (c) and 35S::FK (d). Bars 2 cm. e FK mRNA levels in 15-day-old Col-0, fk-J3158, ProFK::FK/fk-J3158 and 35S::FK

Full size image
Fig. 2
figure 2

Phenotypes of flowering in fk-J3158, ProFK::FK/fk-J3158 and 35S::FK under different photoperiods. 40-day-old plants were grown under LD conditions (16 h light/8 h dark) (a), and 55-day-old plants were grown under SD conditions (8 h light/16 h dark) (b). From left to right were Col-0, fk-J3158, ProFK::FK/fk-J3158 and 35S::FK plants. Bars 10 cm. Flowering time was indicated as bolting days (c) and rosette leaves (d) under LD conditions (n = 60). Flowering time was indicated as bolting days (e) and rosette leaves (f) under SD conditions (n = 25). Asterisk indicates significant difference from the value of Col-0 (t test, P < 0.05)

Full size image

FK is expressed in tissues with high mitotic activity

The fk-J3158 plants displayed a late-flowering phenotype. Many flowering genes are expressed in tissues with high mitotic activity, especially floral organs, in order to regulate flowering. Previous studies showed that FK is strongly expressed in the shoot apical meristem (Jang et al. 2000; Schrick et al. 2000), but its expression in the floral meristem has not been characterized. In order to further analyze the involvement of FK in flowering time, we characterized the organ- and tissue-specific of FK expression patterns using ProFK::GUS plants. We observed FK expresses in the hypocotyl and cotyledons, immature roots, immature leaves, shoot meristem, immature clip, and in the flowers (Fig. 3a–f). In floral organs, the GUS signal was very strong in sepals, anthers, stigma, carpel, and pollen, and absent in petals (Fig. 3e). These results suggested that FK is expressed in both the shoot meristem and floral organs to regulate flowering.

Fig. 3
figure 3

Expression pattern analysis of FK. a and f Translational fusions of the FK regulatory sequences with the GUS reporter gene were examined in transgenic Arabidopsis plants. Histochemical staining was performed on 3-day-old seedlings (a), 25-day-old plant (b), shoot (c), root (d), floral organ (e), and silique (f). Bars 1 mm

Full size image

BRs are unable to specifically complement the late-flowering phenotype in fk-J3158

Sterols are a class of membrane isoprenoid lipids that are essential to regulate embryogenesis and post-embryonic stages of plant development via a currently unidentified signaling pathway in Arabidopsis (He et al. 2003; Lindsey et al. 2003; Men et al. 2008; Babiychuk et al. 2008; Qian et al. 2013). It is known that FK functions upstream of BR biosynthesis. Previous studies reported that BRs are important signals to promote the floral transition by repressing the expression of FLC (Chory et al. 1991; Azpiroz et al. 1998; Turk et al. 2005; Domagalska et al. 2007). To determine whether specific BRs could rescue the late-flowering phenotype of fk mutant plants, seeds of fk-J3158 were sown on media containing in 1, 10, and 100 nM of 24-epibrassinolide (BL) (Fig. 4a, b). When fk-J3158 was treated with 1 and 10 nM of BL, the frequency of bolting plants in 27 day were significantly enhanced, suggesting a partial rescue of the late-flowering phenotype in the mutant. However, this rescued flowering phenotype was repressed under the treatment with 100 nM of BL. Moreover, these results were very similar to those seen in the BL-treated Col (Fig. 4a, b; Suppl. Fig. S1). These data demonstrated that FK, which functions in the early sterol biosynthetic pathway to regulate flowering, may be independent of the BR-signaling pathway, but the final product BR can indirectly affect the flowering of early-step sterol biosynthetic mutants.

Fig. 4
figure 4

Dose-response of fk-J3158 flowering time to different concentrations of BL treatment and expression levels of autonomous pathway regulators. Twenty-five-day-old seedlings of Col-0 and fk-J3158 were grown on MS plate after BL treatment under LD conditions and the frequency of bolting (a) and days to bolting (b) were measured (n = 40). c Transcript levels of FLK, LD, FVE, and FLD in 15-day-old seedlings

Full size image

Expressions of a number of autonomous pathway regulators are not affected in fk-J3158

It is known that FLK, LD, FVE, and FLD are important regulators of FLC in the autonomous pathway. Since the expression of FLC was up-regulated in fk-J3158, the question was posed whether FK regulation of flowering time is dependent on the autonomous pathway? For this, we investigated the expression of FLK, LD, FVE, and FLD in the fk-J3158 mutant by qRT-PCR analysis under the LD conditions. We found that the expression levels of these genes were not significantly changed (Fig. 4c). This result indicated that FK functioning in the early sterol biosynthetic pathway to regulate flowering may be independent of the autonomous pathway.

The late-flowering phenotype of fk-J3158 is significantly rescued by application of exogenous GA3

In Arabidopsis, genetic and pharmacological experiments implicate GAs as promoters of flowering, particularly under non-inductive SD conditions (Wilson et al. 1992; Jacobsen and Olszewski 1993; Wilson and Somerville 1995). To determine whether the late-flowering phenotype of FK mutants could be rescued by exogenous GA treatment, seeds of fk-J3158 were sown on MS media containing 1, 10, 50, or 100 µM of GA3. Under both LD and SD conditions, exogenous application of GA accelerated flowering in Col plants (Fig. 5a–c; Suppl. Fig. S2). Interestingly, fk-J3158 was more sensitive to GA3 than Col. Under LD conditions, the bolting days of fk-J3158 had been rescued to the wild-type level under 1 µM GA3 treatment, and even lower than the Col (Fig. 5a–c). Similar phenotypes were also evident under SD conditions (Suppl. Fig. S2). Concurrently, the expression of FLC in fk-J3158 following GA3 treatment was rescued to the untreated-GA3 wild-type level. However, we observed only a little difference in Col following GA3 treatment (Fig. 5d). To further demonstrate that the delayed flowering phenotype of fk-J3158 was rescued by GA3, we sprayed fk-J3158 plants grown under LD conditions with a 100 µM GA3 solution. Comparable to the results observed following GA3 treatment on MS media, the days to bolting and total leaf number were also significantly reduced than those in Col (Fig. 5e–h). In addition, when plants grown under LD conditions were sprayed with 37 mg/l paclobutrazol (PAC; the GA biosynthetic inhibitor) solution, both the Col and fk-J3158 plants displayed similar degree of increased bolting days and total leaf number (Fig. 5e–h). These results suggested that the role of FK in the regulation of flowering may be connected to the GA pathway.

Fig. 5
figure 5

Response of fk-J3158 flowering time to GA3 and PAC treatment under LD conditions. Twenty-five-day-old seedlings of Col-0 (a) and fk-J3158 (b) were grown on MS plate after different concentrations of GA3 treatment under LD conditions. Bars 2 cm. c Measurement of days after bolting under LD conditions (n = 40). d Transcript levels of FLC in 14-day-old seedlings after GA3 treatment under LD conditions. Fifty-day-old of Col-0 (e) and fk-J3158 (f) plant were grown under LD conditions in soil. Plants were sprayed with 100 μM GA3 twice a week (+GA3) or without GA3 (CK) or with 37 mg/l PAC treatment once a week (+PAC). Bars 10 cm. g, h Measurement of days after bolting and rosette leaf number under GA3 and PAC treatment. Values represent the mean ± standard deviation (n = 20)

Full size image

FK function and the GA pathway interact with each other in floral transition

Since FK was found to be involved in GA-induced flowering, we wanted to know how FK affected the GA pathway. The expression levels of GA20ox1, GA3ox1 and GA2ox2 were determined in 15-day-old Col and fk-J3158 seedling by qRT-PCR. Expression levels of GA20ox1 and GA3ox1 were down-regulated in fk-J3158, GA2ox2 expression was up-regulated (Fig. 6a). It is known that GA20ox1 and GA3ox1 promote GA biosynthesis and GA2ox2 assists in GA degradation. Thus, these data suggested that FK mutation may lead to an absence of endogenous GAs and an increase in DELLA in fk-J3158. A recent report has shown that DELLA proteins interact with FLC to enhance the inhibition ability of FLC (Li et al. 2016). Moreover, FK expression was also detected following GA3 and PAC treatment. We found that FK expression was enhanced following GA application but reduced following PAC treatment (Fig. 6b), which is consistent with the results in a previous study (Jang et al. 2000). In the double-mutant analysis, we observed that ga1-3 could enhance the fk-J3158 seedling phenotype (Fig. 6c). In addition, no additive effect was observed in the fk-J3158 ga1-3 double mutant under LD conditions (Fig. 6c–e) and it did not flower during the extended growth period as was observed for ga1-3. These results could suggest that GA1 is potentially epistatic to FK in the regulation of flowering. Together, these results suggested that FK function and the GA pathway may interact to regulate floral transition.

Fig. 6
figure 6

FK interacts with gibberellin pathway in flowering. a Transcript levels of GA20ox1, GA3ox1, GA2ox2 and GID1 in 14-day-old fk-J3158 seedlings under LD conditions. b Transcript levels of FK in 14-day-old wild-type plant by gibberellins and PAC treatment under LD conditions. Asterisk indicates significant difference from the value of Col-0 (t test, P < 0.05, n = 5). c Floral-timing phenotypes of 35-day-old Col-0, Ler-0, fk-J3158, ga1-3, Col-0 Ler-0, and fk-J3158 ga1-3 under LD conditions. Bars 4 cm. Total leaf number at bolting (d) and days to bolting (e) was analyzed in 35-day-old Col-0, Ler-0, fk-J3158, ga1-3, Col-0 Ler-0, and fk-J3158 ga1-3

Full size image

The late-flowering phenotype of fk-J3158 is significantly suppressed by vernalization treatment

Vernalization is an important process that promotes flowering in response to prolonged exposure to low temperature (Gendall et al. 2001; Levy 2002; Sung and Amasino 2004). To determine whether vernalization can prevent the late-flowering phenotype of fk-J3158, the Col and fk-J3158 plants were grown at 4 °C for 6 weeks before they were transferred to 23 °C under LD conditions. Under vernalization treatment, the average total leaf number of Col plants was reduced from 12.11 to 10.52. However, the average total leaf number of vernalization-treated fk-J3158 plants was reduced from 18.34 to 11.44, it was much earlier than the untreated plants. Vernalization-treated fk-J3158 plants also had a lower average total leaf number than that of untreated Col plants (Fig. 7a). Furthermore, the expression of FLC was also analyzed because vernalization is known to repress FLC transcription. We found that the expression level of FLC was high in the untreated fk-J3158 plants but was drastically decreased after vernalization (Fig. 7b), which was consistent with the data concerning leaf number. Thus, these results suggested that the role of FK in the regulation of flowering may involve the vernalization pathway.

Fig. 7
figure 7

Leaf number and FLC transcript level in fk-J3158 under vernalization. a Measurement of total leaf number for Col-0 and fk-J3158 mutants under cold-treated (3 weeks) (+Vrn) or untreated (−Vrn). Values represent the mean ± standard deviation (n = 20). b FLC expression in fk-J3158 mutant after vernalization treatment. Values represent the mean ± standard deviation (n = 5)

Full size image

FK function in regulating flowering is partially dependent on FLC

It is known that several flowering regulation pathways, including the autonomous pathway, vernalization pathway, and BR signaling, converge on the transcriptional regulator FLC that represses the floral transition (Michaels and Amasino 1999; Sheldon et al. 1999; Michaels and Amasino 2001; Domagalska et al. 2007). Many late-flowering mutants have been identified via genetic screens based on high-level FLC expression. Thus, we investigated FLC expression in the fk-J3158 mutant using qRT-PCR analysis of plants grown under the LD conditions. As expected, we detected low-level FLC expression in Col, but high-level of FLC expression in fk-J3158. Moreover, FLC expression was reduced in ProFK::FK/fk-J3158 and 35S::FK plants (Fig. 8a). FLC expression levels were consistent with the lifespan of fk-J3158, ProFK::FK/fk-J3158 and 35S::FK plants (Fig. 8a). FLC can repress SOC1 expression by directly binding to specific regulatory elements (Searle et al. 2006). Corresponding to FLC expression levels, SOC1 expression was reduced in fk-J3158 and enhanced in ProFK::FK/fk-J3158 and 35S::FK plants (Fig. 8b). The previous experimental, the late-flowering phenotype of fk-J3158 is significantly rescued by exogenous GA3 application and vernalization treatment. The expression level of FLC was reduced by these same treatments. GA can also promote the degradation of DELLA to reduce the inhibition ability of FLC, which leads to early flowering. To further determine which the late-flowering phenotype in fk-J3158 was related to FLC, we constructed the double mutant fk-J3158 flc to confirm whether FLC mutation can rescue the late-flowering phenotype of fk-J3158 (Fig. 9a). It was found that the flowering time of double mutant fk-J3158 flc was significantly earlier than fk-J3158 but later than Col under the LD conditions (Fig. 9a, b). This result indicates that FK regulation of flowering is partially dependent on FLC.

Fig. 8
figure 8

Analysis of flowering genes FLC and SOC1 in Col, fk-J3158, ProFK::FK/fk-J3158 and 35S::FK seedlings under LD and SD conditions. a Transcript levels of FLC in 15-day-old seedlings under LD and SD conditions. b Transcript levels of SOC1 in 15-day-old seedlings under LD and SD conditions. Asterisk indicates significant difference from the value of Col-0 (t test, P < 0.05, n = 5)

Full size image
Fig. 9
figure 9

The phenotype of double mutant fk-J3158 flc. a The phenotype in 40-day-old Col, flc, fk-J3158, fk-J3158 flc under LD conditions. Bars 10 cm. Flowering time was measured as bolting days (b) and rosette leaves (c) under LD conditions (n = 30). Asterisk indicates significant difference from the value of Col-0 (t test, P < 0.05)

Full size image

Discussion

The sterol biosynthetic pathway in Arabidopsis has been well described through the identification and characterization of mutants compromised in each biosynthetic step. Previous studies indicate that the sterol biosynthetic pathway plays a role in many different processes including morphogenesis, cell differentiation, cell polarity, and cell patterning (Lindsey et al. 2003; Schaller 2003; Boutté and Grebe 2009). BRs are synthesized from the sterols campesterol. It was reported that most BR-biosynthetic mutants and BR-insensitive mutants exhibited a delayed flowering time. The majority of mutants compromised in the sterol biosynthetic pathway show reduced endogenous BR levels and also display late-flowering phenotypes. Because the mutants compromised in early-step sterol biosynthesis display severe multiple defects in plant growth that cannot be rescued by exogenous BR application, it is unknown whether the late-flowering phenotypes in these sterol mutants are dependent on the BR pathway. In this study, the weak FK mutant allele fk-J3158 resulted in plants that displayed late flowering and that were not rescued by exogenous BL treatment, these results suggest that FK or early-step of sterol biosynthesis is involved in an additional pathway that regulates floral transition, which is independent of sterols and BRs. Many studies have suggested that there are potentially a number of unidentified intermediates in the sterol biosynthetic pathway that mediate a novel signaling pathway to regulate the development of embryos, vascular tissue, and stomatal development in Arabidopsis, in a manner independently of sterols and BRs (Jang et al. 2000; Schrick et al. 2000, 2002; Souter et al. 2002; Kim et al. 2005; Men et al. 2008; Qian et al. 2013). It is possible that the late-flowering phenotypes of early-step sterol biosynthetic mutants are the result of disruptions in the synthesis of these unidentified intermediate molecules.

Several floral-transition signaling pathways have been characterized, including the photoperiod, autonomous, vernalization, BR, and GA pathways. The fk-J3158 plants were sensitive to variation in photoperiod like the wild type and expression of a number of genes involved in the autonomous pathway were not significantly changed in the mutant plants, suggesting that FK might be not be involved in the photoperiod and autonomous pathway to regulate flowering. Several pathways are activated by epigenetic silencing of FLC (Simpson and Dean 2002; Sung and Amasino 2004; Alexndre and Hennig 2008), which can act as a negative regulator of SOC1. We found that FLC expression was enhanced and SOC1 expression was reduced in fk-J3158. Moreover, following exogenous GAs treatment, the late-flowering phenotype was significantly rescued in fk-J3158 and the expression level of FLC was also reduced to the wild-type level. In addition, the late-flowering phenotype and the FLC expression level in fk-J3158 were also rescued by vernalization treatment. We demonstrated that flowering time in the double mutant fk-J3158 flc is significantly earlier than that in fk-J3158 under the LD conditions. This result indicates that FK regulating flowering is partially dependent on the FLC, and that FK regulates floral transition mainly via FLC that integrates the GA and vernalization pathways.

In a previous study, the GA pathway was found to directly regulate SOC1 expression, but this is a function in the downstream of FLC because GAs did not influence the expression of FLC. In a recent study, DELLA was shown to interact with FLC to enhance the inhibition ability of FLC. The inhibition ability of FLC is reduced when exogenous GA promoted the degradation of DELLA (Li et al.2016). This may be one of main reasons that GA promotes flowering. However, in this study, FLC expression in fk-J3158 plants was partially rescued by GA treatment. In other studies, it was reported that overexpression of GASA5 which is involved in GA signaling, also resulted in high-level FLC expression. Moreover, the low expression levels of GA biosynthetic genes suggest that endogenous GAs are absent in fk-J3158 plants. Based on the fact that the FK expression can be regulated by exogenous GAs and PAC, it was suggested that the FK and GA pathway interact to regulate flowering time and that GA may affect FLC expression indirectly. It is known that hormone and environment interaction exists in the regulation of seed germination, in which the levels of GAs can be up-regulated by vernalization to break dormancy and promote germination (Yamauchi et al. 2004). In addition, FLC also play an important role in GA and temperature-dependent germination (Chiang et al. 2009). In the present study, the late-flowering phenotype of fk-J3158 can be attributed to the high expression level of FLC and the low level of GA. Vernalization treatment and GA application can reduce the expression level of FLC in fk-J3158. This result indicates that FK regulating of FLC expression is dependent on both the vernalization and GA pathways. Exogenous GA application also can promote DELLA degradation that reduces the suppression ability of FLC. These factors together rescue the late-flowering phenotype of fk-J3158 plants. We also demonstrated that the double mutant fk-J3158 flc can rescue the late-flowering phenotype of fk-J3158. This result indicates that FK’s regulation of flowering is partially dependent on FLC. Therefore, it may be possible that vernalization and GA pathways interact with each other thus affecting floral transition indirectly by regulating FLC. Although it was reported that other mutants in the early steps of the sterol biosynthetic pathway also display a late-flowering phenotype, it would be interesting to test whether these defects could be rescued by GA and/or vernalization treatments.

Author contribution statement

BH, PQ, NG and JS conducted experiments and analyzed data. SH, PQ and BH designed the research and wrote the manuscript. All authors read the manuscript and approved the manuscript.