Flowering Time Modulation by a Vacuolar SNARE via FLOWERING LOCUS C in Arabidopsis thaliana

The transition of plant growth from vegetative to reproductive phases is one of the most important and dramatic events during the plant life cycle. In Arabidopsis thaliana, flowering promotion involves at least four genetically defined regulatory pathways, including the photoperiod-dependent, vernalization-dependent, gibberellin-dependent, and autonomous promotion pathways. Among these regulatory pathways, the vernalization-dependent and autonomous pathways are integrated by the expression of FLOWERING LOCUS C (FLC), a negative regulator of flowering; however, the upstream regulation of this locus has not been fully understood. The SYP22 gene encodes a vacuolar SNARE protein that acts in vacuolar and endocytic trafficking pathways. Loss of SYP22 function was reported to lead to late flowering in A. thaliana plants, but the mechanism has remained completely unknown. In this study, we demonstrated that the late flowering phenotype of syp22 was due to elevated expression of FLC caused by impairment of the autonomous pathway. In addition, we investigated the DOC1/BIG pathway, which is also suggested to regulate vacuolar/endosomal trafficking. We found that elevated levels of FLC transcripts accumulated in the doc1-1 mutant, and that syp22 phenotypes were exaggerated with a double syp22 doc1-1 mutation. We further demonstrated that the elevated expression of FLC was suppressed by ara6-1, a mutation in the gene encoding plant-unique Rab GTPase involved in endosomal trafficking. Our results indicated that vacuolar and/or endocytic trafficking is involved in the FLC regulation of flowering time in A. thaliana.


Introduction
Membrane trafficking is a major regulatory system for protein transport in eukaryotic cells. Evolutionary conserved regulatory molecules play substantial roles in regulating the budding of transport vesicles from organelles and the fusion to target membranes. For example, SAR/ARF GTPases regulate budding processes, and RAB GTPases and soluble N-ethyl-maleimide sensitive factor attachment protein receptors (SNAREs) are key regulators of membrane tethering and fusion. The SNARE proteins contain a heptad-repeat called the SNARE motif; according to sequence similarities in SNARE motifs, they are classified into three Q-SNARE groups and one R-SNARE group. In most cases, Q-SNAREs localize on target organelle membranes, and R-SNAREs localize on transport carriers. Typically, four SNARE molecules (or three, in the complex that contains a SNAP25-type SNARE) form a tight, stable complex that promotes membrane fusion between transport carriers and target organelles [1]. It is widely accepted that the combination of four SNARE proteins, in coordination with a particular RAB GTPase, confers specificity in membrane fusion.
SYP22/VAM3/SGR3 is a Q-SNARE that acts in vacuolar and endocytic transport pathways; this gene was originally identified by functional complementation of the yeast vam3 mutant [2]. SYP22 localizes on the endosome/prevacuolar compartment (PVC) and the vacuolar membrane [2,3,4,5]. There, SYP22 forms a complex with VTI11/ZIG, SYP5, and VAMP727 [6,7,8]. Loss-of-function SYP22 mutants exhibit a wide spectrum of phenotypes, including wavy leaves, semi-dwarfism, immature leaf vascular tissues, excessive differentiation of idioblasts, and late flowering [9,10,11]. These phenotypes were synergistically enhanced by a second mutation in the cognate SNARE gene, VAMP727; conversely, the overexpression of VAMP727 completely suppressed all syp22-1 phenotypes [6]. Thus, syp22-1 phenotypes appeared to result from defects in the endocytic and/or vacuolar transport pathways mediated by the SNARE complex of SYP22 and VAMP727, and not from defects in other unidentified SYP22 functions that are independent of membrane trafficking.
Most phenotypes of syp22 mutants, like wavy leaves, semidwarfism, and abnormal leaf vascular tissues, are thought to be caused by a defect in polar auxin transport; the syp22-4 mutation affects the polarized localization of PIN1, an efflux carrier of auxin, and the polar transport of auxin [10]. However, no studies have reported crosstalk between polar auxin transport and flowering regulation; this suggested that the late flowering phenotype of syp22 could be due to a defect in an uncharacterized mechanism that linked flowering regulation to vacuolar/endocytic trafficking. In this study, we examined the mechanism underlying the late flowering phenotype of syp22 to unveil this hidden link.
Upon floral induction, plants in the vegetative stage start producing floral meristems around the shoot apical meristem (SAM). In A. thaliana, the FT gene encodes an important floral inducer that is synthesized in leaves and transported to the SAM to induce expression of other floral inducers, including the SUP-PRESSOR OF OVEREXPRESSION OF CONSTANS1 (SOC1). Expression of FT and SOC1 in A. thaliana is known to be regulated by at least four distinguishable genetic pathways: the photoperiodic-dependent, gibberellin-dependent, vernalization-dependent, and autonomous promotion pathways [12,13]. The photoperiodic pathway modulates expression of FT in leaves in response to changes in the photoperiod [14]. The gibberellin pathway regulates flowering by controlling expression of SOC1 in the SAM [15]. The vernalization and autonomous pathways cooperatively modulate expression of FLC, which represses the expression of FT and SOC1 by directly interacting with the promoter regions of these genes [16]. Thus, FLC is a key integrator of the vernalization-dependent and autonomous promotion pathways.
To date, many putative nuclear factors have been identified that act in the autonomous pathway, including the RNA-binding or RNA-processing proteins, FCA, FPA, FLK, and FY, and chromatin modifiers, FVE and FLD [13]. However, no studies have reported the involvement of membrane trafficking in the regulation of FLC expression. Our results highlighted an unexpected link between the vacuolar SNARE and autonomous regulation of flowering in A. thaliana.

syp22-1 was Defective in Determination of Floral Meristem Identity
Previously, syp22-1 was reported to exhibit a broad range of phenotypes, including weak late flowering under continuous light (CL) (Fig. 1A) [9]. First, we investigated whether there was a functional link between SYP22 and floral meristem identity determination. Under normal growth conditions (CL, 23uC), determination of the floral meristem identity was not affected in syp22-1 (Fig. 1B). However, when grown under CL at 16uC, or under short-day conditions (SD: 8 h light and 16 h dark) at 23uC, the floral meristems of syp22-1 were frequently converted to inflorescence meristems (Fig. 1C, D), and aerial rosettes were occasionally produced (Fig. 1E). A similar phenotype has been reported for other late flowering mutants, including fld-2, a mutant defective in the autonomous pathway [17].
We then examined whether the abnormality in determination of floral meristem identity was related to the regulatory mechanism involving LEAFY (LFY), a key determinant of meristem identity. We generated double mutants between syp22-1 and either a weak (lfy-2) or null (lfy-1) allele. The syp22-1 mutation exaggerated the lfy-2 phenotype, which suggested that the two genes operated in a synergistic manner; the double mutant syp22-1 lfy-2 exhibited an indistinguishable phenotype from lfy-1, even when the double mutant plants were grown at 23uC in CL ( Fig. 2A). On the other hand, the double mutant syp22-1 lfy-1 exhibited the same phenotype as lfy-1. Similar synergistic enhancement of weak alleles of lfy mutations was reported in double mutants of lfy combined with other late flowering mutations, ft, fd, and fld [17,18,19].
We then analyzed the morphology of 15 proximal axillary organs that formed along the stems of wild-type and mutant plants. Compared with wild type and syp22-1, the lfy-2 plants generated a larger number of lateral shoots, with or without a bract, instead of flowers. This was also synergistically enhanced in the double mutant, syp22-1 lfy-2 ( Fig. 2B, p,0.01, Student's t-test). These results indicated that impairment of SYP22 function caused a deleterious effect on the determination of meristem identity through a pathway that involved LFY.

syp22-1 Responded Normally to Gibberellin and Vernalization
The four genetically distinguishable pathways that regulate the flowering time of A. thaliana are integrated by the expression of flowering pathway integrators, including the LFY gene [20]. To examine whether the syp22-1 mutant was defective in these regulatory pathways, we first tested the syp22-1 response to gibberellic acid (GA 3 ). We found that GA 3 treatment significantly promoted the flowering of syp22-1 ( Fig. 3A and Fig. S1A, p,0.01, Student's t-test); this indicated that syp22-1 retained responsiveness to gibberellin. Thus, the late flowering phenotype of syp22-1 did not seem to be caused by impairment in the gibberellin-dependent pathway.
Next, we tested the syp22-1 response to the photoperiod. The wild-type plants (Columbia accession) used in this study flowered after generating 12.161.1 leaves under CL conditions and after generating 42.7614.8 leaves under SD. On the other hand, the syp22-1 mutant generated a significantly larger number of leaves (79.8623.4 leaves) when grown under SD than the wild type plants grown under SD ( Fig. 3B and Fig. S1B, p,0.01, Student's t-test). In contrast, the late flowering mutants defective in the photoperiod-dependent pathway are known to be insensitive to changes in photoperiod [21]; thus, the photoperiodic pathway did not seem to be markedly affected by the syp22-1 mutation.
We then demonstrated that syp22-1 responded normally to vernalization; vernalization treatment for 4 weeks at 4uC promoted flowering of syp22-1 under SD (Fig. 3B, p,0.01, Student's ttest). These results indicated that SYP22 was involved in the regulation of flowering time through regulatory pathways other than the above three pathways. Furthermore, we confirmed that the late flowering phenotype could be rescued by introducing the GFP-SYP22 chimeric gene into the syp22-1 mutant, under the regulation of its own promoter (Fig. S2); this confirmed that the late flowering phenotype was due to a loss of function in SYP22.

Expression Level of FLC was Elevated in syp22-1
The results mentioned above strongly suggested that the late flowering of syp22-1 was caused by impairment in the last of a series of major regulatory pathways, the autonomous promotion pathway. FLC is a key regulator of the autonomous pathway. FLC encodes a MADS transcription factor that represses the expression of the floral pathway integrators, FT, SOC1, and LFY [13]. Thus, we reasoned that, if delayed flowering of syp22-1 was due to a defect in autonomous flowering promotion, then the mRNA level of FLC might be altered in this mutant. Accordingly, we examined the expression level of FLC and the floral pathway integrators in 14-day-old seedlings of the syp22-1 mutant by quantitative RT-PCR (qRT-PCR). In syp22-1, FLC mRNA expression exceeded that of wild type by about three-fold (Fig. 4, p,0.01, Student's ttest). Moreover, the mRNA expression of FT was significantly reduced compared to wild type (Fig. 4, p,0.01, Student's t-test). In the syp22-1 mutant transformed with GFP-SYP22, accumulation of mRNA for FLC was comparable to that in wild type (Fig. S2C). These results strongly suggested that the late flowering phenotype of syp22-1 could be ascribed to elevated expression of FLC. We also confirmed this genetically; we reasoned that, if the elevated expression of FLC mRNA in syp22-1 had directly caused its late flowering phenotype, then the elimination of FLC activity should suppress this phenotype. To test this, we generated a double mutant of syp22-1 and flc-3, a null allele of flc, in a Columbia background. The endogenous expression of FLC was low in the Columbia accession, and flc-3 exhibited a phenotype similar to wild type when grown in CL (Fig. 5) [22]. On the other hand, as expected, the flc-3 syp22-1 double mutant generated a significantly smaller number of leaves (14.263.4 leaves) before flowering than syp22-1 (23.763.9 leaves, p,0.01, Student's t-test) under CL conditions; thus, the double mutant was comparable to the wild type (12.361.1 leaves) and flc-3 (11.461.7 leaves) plants (Fig. 5). These results clearly demonstrated that the elevated level of FLC mRNA expression conferred a late flowering phenotype on syp22-1. Among the pleiotropic phenotypes of syp22-1, the flc-3 mutation suppressed only the late flowering phenotype; in contrast, other phenotypes, like wavy leaves and semi-dwarfism, were not remedied in the flc-3 syp22-1 double mutant (Fig. 5A). This indicated that the late flowering phenotype of syp22-1 was conferred by a mechanism distinct from that responsible for other phenotypes, which are most likely due to impairments in polar auxin transport.

Mutations in BIG, a Putative Regulator of Membrane
Trafficking, and ARA6 also Altered FLC Expression The results described above strongly suggested that vacuolar and/or endocytic trafficking pathways were involved in autonomous regulation of flowering. We then examined whether other membrane trafficking mutants exhibited a similar phenotype. We tested whether a genetic mutant of BIG, which encodes a Calossinlike protein, might also exhibit elevated FLC expression. BIG was originally identified as the gene responsible for the doc1 mutation [23]. BIG was also reported to be required for the auxin-mediated inhibition of endocytosis [24]. Furthermore, treatment with naphthylphthalamic acid (NPA), an inhibitor of polar auxin transport, altered subcellular localization of PIN1 in the doc1 mutant [23]. Based on these results, mutations in BIG were implicated in impairments of membrane trafficking pathways, including the endocytic and/or vacuolar transport pathways. Intriguingly, some mutant alleles of BIG have been also reported to exhibit the late flowering phenotype [25,26]. These results implied that BIG was also involved in regulation of FLC expression. Among the big mutations, the allele that caused the most severe late flowering phenotype was doc1-1 [26]. Under our CL conditions, doc1-1 exhibited a late flowering phenotype (13.961.2 leaves, Fig. 6A, C) compared with wild type (12.060.8 leaves); this phenotype was similar to that of syp22-1 (16.361.7 leaves, Fig. 6A, C). We then generated a doc1-1 syp22-1 double mutant, which exhibited a more severe late flowering phenotype (21.063.6 leaves, Fig. 6A-C, p,0.01, Student's t-test). A qRT-PCR analysis also indicated that the level of FLC mRNA expression was slightly increased in doc1-1, and this was enhanced in the double mutant, doc1-1 syp22-1 (Fig. 6C). Thus, BIG and SYP22 could be involved in the same regulatory pathway for FLC expression.
We then examined the effect of a mutation in the other membrane trafficking regulator, ARA6, which is a plant-unique Rab GTPase involved in trafficking between multivesicular endosomes and the plasma membrane [27]. We have already reported that the ara6-1 mutation suppressed the late flowering phenotype of syp22-1. We examined whether the elevated expression level of FLC in syp22-1 is also suppressed by ara6-1, and found that FLC expression was significantly reduced in ara6-1 syp22-1 (Fig. 6D, p,0.05, Student's t-test). This result also

Discussion
The syp22 mutants have exhibited pleiotropic phenotypes, including late flowering, semi-dwarfism, waved leaves, immature leaf vascular tissues, and excessive idioblast differentiation. All these phenotypes, except late flowering, have been thought to result from a misdistribution of the phytohormone, auxin [9,10,11]. In contrast, the reason that mutations in SYP22 engendered late flowering remained unclear. In the present study, we demonstrated that elevated expression of FLC in the syp22-1 mutant was a direct cause of the late flowering phenotype (Figs. 4 and 5). The syp22-1 mutant was capable of responding to vernalization treatment (Fig. 3D), which suggested that the syp22-1 mutation resulted in impairment in the autonomous promotion pathway. We further demonstrated that the expression level of FLC was elevated in another membrane trafficking mutant, doc1-1 (Fig. 6A-C), and that ara6-1, a mutation in the gene encoding a plant-unique Rab GTPase, suppressed elevated expression of FLC in syp22-1 (Fig. 6D). These results indicated that endocytic and/or vacuolar trafficking pathways were involved in the autonomous promotion of flowering in the Columbia accession of A. thaliana.
Our findings clearly indicated that the vacuolar SNARE, SYP22, was required for proper regulation of FLC, a key regulator of flowering time. This result may elucidate the recent finding that TERMINAL FLOWER 1 (TFL1), a homolog of FT, which is required for repression of the transition from the inflorescent meristem to the floral meristem, played a critical role in membrane trafficking to protein storage vacuoles (PSVs) [28]. Based on that finding, the authors proposed a novel function of the  PSV: the storage of factors necessary for flowering and meristem maintenance. Other genetic studies indicated that TFL1 acted downstream of FVE and FCA, which are upstream regulators of FLC in the autonomous pathway [19,29]. In addition, tfl1-1, a lossof-function mutant of TFL1, exhibited an early flowering phenotype and reduced expression of FLC [30]. Those results suggest that TFL1 may also be involved in autonomous promotion of flowering. In our previous study, we found that the syp22 mutant also exhibited defects in the biogenesis of PSVs in A. thaliana [6]. Taken together, these results suggest the hypothesis that syp22-1 is defective in transport or storage of the flowering factors harbored in PSVs that affect autonomous promotion of flowering via modulation of FLC expression. Our present results clearly indicated that the flowering phenotype of syp22-1 was caused by a mechanism genetically separable from that responsible for other phenotypes, like semi-dwarfism and wavy leaves, which are most likely due to abnormal polar auxin transport. However, the predicted flowering regulators could be transported via the same trafficking pathway as the efflux carriers of auxin, because mutants of both SYP22 and BIG/DOC1 also exhibited abnormalities in polarized localization of the PIN1 protein [10,23].
Another potential hypothesis is that SYP22 could be involved in transcriptional regulation of FLC via chromatin remodeling. In mammalian cells, an adaptor protein containing PH domain, PTB domain, and leucine zipper motif 1 (APPL1) has been identified as a RAB5 effector, which transmits a signal from the endosomes to the nucleus [31]. APPL1 was also shown to interact with the nucleosome remodeling and deacetylase (NuRD) complex to regulate histone deacetylase (HDAC) 2 activity [32]. The level of FLC expression is tightly regulated, and this depends on the extent of chromatin modification; moreover, FVE and FLD were demonstrated to be components of the HDAC complex responsible for deacetylation of the FLC locus [33,34]. If similar regulatory mechanisms of endosomal signaling followed by chromatin modification are conserved in plants, the mutation in SYP22 or BIG could affect these machineries, which could lead to elevation of FLC expression. Further study of molecular components that mediate endosomal signaling in plant cells would be required to investigate these hypotheses.
In conclusion, this study revealed a linkage between the fundamental cellular function of membrane trafficking and the higher ordered regulation of plant flowering. Future study of flowering from the viewpoint of organelle functions that are associated with membrane trafficking would elucidate our understanding of this regulatory system.

Plant Materials and Growth Conditions
A. thaliana (Columbia accession) seeds were sown on MS plates after sterilization. After a 2-day incubation at 4uC, plants were grown at 23uC under CL or SD (8 h light, 16 h dark). After 14 days (CL) or 21 days (SD) of incubation on plates, plants were transferred to the soil. For the vernalization treatment, plants were grown for 3 weeks at 23uC in SD, then incubated at 4uC for 8 weeks in SD.

GA 3 Treatment
Seeds were sown on MS plates and cultured for 14 days under CL conditions. Plants were transferred to the soil, and then sprayed with 20 mM GA 3 (Wako) or water twice per week.