Introduction

The transition from vegetative growth (producing stems and leaves) to reproductive development (producing flowers) is a major and complex process in the life cycle of flowering plants, which was widely regulated by multiple environmental aspects and endogenous signals (Higgins et al. 2010; Abou-Elwafa et al. 2011).

There is an extensive progress in the gene and molecular mechanisms of flowering controlling process recently in Arabidopsis (Srikanth and Schmid 2011; Zhang et al. 2014) and varieties of crop species (Kim et al. 2009). In Arabidopsis, which requires for the condition of vernalization and long days facultatively, several regulatory pathways have been defined that is different in their response to distinct internal and external cues, including vernalization, photoperiod, gibberellin (GA), temperature, age and autonomous pathways (He and Amasino 2005; Michaels 2009; Srikanth and Schmid 2011; Zhang et al. 2014). The autonomous pathway functions promote the flowering independently of day length by repressing the central flowering repressor and vernalization regulatory gene FLOWERING LOCUS C (FLC) (also termed as AGAMOUS-LIKE 25/AGL25, FLOWERING LOCUS F/FLF, REDUCED STEM BRANCHING 6/RSB6) (Michaels et al. 2003; Pazhouhandeh et al. 2011; Abou-Elwafa et al. 2011; Yan et al. 2010). Comparisons between Arabidopsis and rice have revealed that rice contains several homologues of many of the known flowering time regulatory genes, and moreover, that certain aspects of the photoperiodic and autonomous flowering pathways are well conserved in this species (Lagercrantz 2009). Understanding the regulation of the floral transition in different plant species provides with important perceptiveness for the possible ancestral control of flowering and the alternatively evolutional mechanisms in different plants (Higgins et al. 2010).

Based on protein sequence similarity, FLC belongs to a six-gene sub-family of the MADS-box class of transcription factors. Through inhibiting the expression of the floral primordium identity genes FLOWERING LOCUS T (FT), SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1/AGAMOUS-LIKE 20 (SOC1/AGL20), LEAFY (LFY), APETALA 1 (AP1) and floral organ identity genes AGAMOUS (AG) and AP3, FLC hinders the floral transition at a quantitative style (Boss et al. 2004; Kobayashi and Weigel 2007; Michaels 2009; Pazhouhandeh et al. 2011).

Autonomous pathway genes repress FLC and thus promote the floral transition. Several genes are known to be involved in this pathway such as FLOWERING LOCUS CA (FCA), FLOWERING LOCUS D (FLD), FLOWERING LOCUS KH DOMAIN (FLK), FLOWERING LOCUS PA (FPA), FLOWERING LOCUS VE (FVE), FLOWERING LOCUS Y (FY), and LUMINIDEPENDENS (LD) (Simpson 2004; Marquardt 2006; Srikanth and Schmid 2011). Most mutants for above genes are recessive and showed late-flowering phenotype under both long- and short-day conditions; however, the related phenotype can be recovered by treatment with vernalization (Abou-Elwafa et al. 2011).

Since the recent reviews about the autonomous flowering pathway (Simpson 2004; Marquardt 2006), considerable progresses in this research field have been made. Hence, evidence now exists to indicate that additional genes including AGL28 (AGAMOUS LIKE 28), CK2 (Casein kinase II), DBP1 (DNA-binding protein phosphatase 1), DRM1 (Developmentally Retarded Mutant1), DRM2 (DOMAINSREARRANGEDMETHYLTRANSFERASE2), ESD4 (EARLY IN SHORT DAYS 4), HDA5 (HISTONE DEACETYLASE 5), HDA6 (HISTONE DEACETYLASE 6), PCFS4 (PCF11P-SIMILAR PROTEIN 4), PEP (PEPPER), PP2A-B’γ (protein phosphatase 2A-B’γ), PRMT5 (protein arginine methyltransferase 5), PRMT10 (protein arginine methyltransferase 10), PRP39-1 (Pre-mRNA Processing Protein 39-1), REF6 (RELATIVE OF EARLY FLOWERING 6), and SYP22 (ARABIDOPSIS THALIANA SYNTAXIN OF PLANTS 22; Table 1) are also likely to be participated in the autonomous flowering time pathway. The present reviews mainly focused on RNA processing and chromatin modification of FLC, post-translational modification of FLC and other mechanisms of FLC regulation in Arabidopsis.

Table 1 Summary of important regulators in autonomous flowering time pathway
Full size table

FCA, FLK, FPA, FY, PCFS4, and PEP are involved in FLC RNA processing

Autonomous pathway genes regulate the levels of FLC mainly through RNA-based post-transcriptional regulation mechanisms and chromatin epigenetic modification (Simpson 2004; Boss et al. 2004; Quesada et al. 2005; Bäurle et al. 2007; Bäurle and Dean 2008). Nine proteins are known to mediate RNA regulating processes: FCA, FLK, FPA, FY, PAPS1, PAPS2, PAPS4, PCFS4 and PEP (Fig. 1).

Fig. 1
figure 1

Six proteins are known to mediate FLC RNA regulatory process: FCA, FLK, FPA, FY, PCFS4 and PEP. FPA interacts with FCA, and both control alternative polyadenylation of antisense RNAs and 3′-end formation at FLC. FCA interacts with the RNA 3′-end processing protein FY. PCFS4 interacts with FCA and regulates the alternative polyadenylation of FCA. These 4 proteins may present in a protein complex and interplay to repress FLC expression via RNA regulatory processes. FLK which is a plant-specific protein with 3 KH RNA-binding domains suppresses FLC expression and its paralogue PEP promotes FLC expression, but the precise molecular mechanism for both needs to be explored further

Full size image

Through genetic dissection of autonomous pathway in Arabidopsis, highly conserved RNA 3′-end processing factors that is plant-specific have been identified (Rataj and Simpson 2014). FCA and FPA, two proteins with a plant-specific RNA-recognition motif-type RNA-binding domain, regulate alternative polyadenylation of antisense RNAs and 3′-end formation at the FLC (Macknight et al. 1997; Schomburg et al. 2001; Liu et al. 2007; Hornyik et al. 2010b; Liu and Mara 2010). FCA is physically and genetically correlated with the RNA 3′-end processing protein FY (Simpson et al. 2003). This interaction is facilitated through 2 proline-rich (PPLPP) motifs in the carboxyl terminal of FY and the WW (typified by two conserved tryptophan residues) domain of FCA and is necessary for both the correct processing of transcripts derived from FCA itself, and the downregulation of FLCexpression (Quesada et al. 2003; Simpson et al. 2003; Srikanth and Schmid 2011). Moreover, FPA and FCA interact with FLD on genetical level, which encodes a histone demethylase, linking RNA processing to chromatin changes in autonomous pathway (Liu et al. 2007). At least the effect of FCA and FPA on levels of FLC and flowering transition partially depends on FLD (Liu et al. 2007; Bäurle and Dean 2008). Hence, FPA and FCA seem to link chromatin-level regulation with RNA processing of FLC (Abou-Elwafa et al. 2011). Functionally, FPA has been linked to FVE in the flowering regulation network. However, it is less sensitive to FLC gene dosage than FVE (Marquardt 2006).

PCFS4 was identified as a new component in the autonomous flowering pathway that regulates the alternative polyadenylation of FCA to promote flowering (Xing et al. 2008). It is hypothesized that PCFS4 interacts with FCA, which regulates alternative RNA processing (Yan et al. 2010), and thus a protein complex consisting of PCFS4, FCA, FPA and FY may regulate FLC expression (Fig. 1).

FLK possesses 3 KH (K-homology) RNA-binding domains and has only been found in plants (Mockler et al. 2004; Lim et al. 2004). The precise molecular mechanism of FLK regulation of FLC expression has yet to be revealed. However, FLK may suppress FLC partially at transcriptional level or through RNA-directed chromatin regulation (Veley and Michaels 2008; Ripoll et al. 2009). Other KH-domain containing proteins in Arabidopsis include RS2-INTERACTING KH PROTEIN (RIK) and HUA ENHANCER 4 (HEN4), which are the part of protein complexes mediating pre-mRNA processing of AG (Cheng et al. 2003).

PEPPER (PEP) is a paralogue of FLK and positively regulates expression of FLC in Arabidopsis (Ripoll et al. 2009). In flk mutant, disruption of PEP can rescue the delayed-flowering transition with a concomitant FLC transcripts decrease. Consistent with this, PEP overexpression resulted in high FLC transcript levels and delayed flowering (Ripoll et al. 2006). The function of FLK and PEP in the autonomous pathway is found to be independent of FCA through genetic and molecular analyses. In addition, PEP may influence FLC RNA level at both transcriptional and post-transcriptional levels. Taken together, PEP is a novel upregulation factor for FLC, emphasizing the importance of RNA-binding activities during flowering transition (Bäurle et al. 2007; Liu et al. 2007; Bäurle and Dean 2008). However, as for FLK, the precise molecular mechanism of PEP in regulating FLC expression needs to be explored further.

Chromatin modification of FLC expression is mediated by DRM2, FLD, FVE, HDA5, HDA6, LD, PRMT5, PRMT10 and REF6

Through chromatin modification, DRM2, FLD, FVE, HDA5, HDA6, LD, PRMT5, PRMT10 and REF6 can regulate FLC expression in Arabidopsis (Fig. 2). Active FLC expression results from a high level of methylated H3K4 around the initiation site of transcription (He and Amasino 2005). FLD, FVE, HDA5, HDA6, LD and REF6 repress FLC expression through histone modification including H3K4 demethylation and H3 or H4 deacetylation in chromatin of FLC (He et al. 2003; Domagalska et al. 2007; Yu et al. 2011; Luo et al. 2015). PRMT5 and PRMT10 play an important role in asymmetric histone arginine methylation to control the floral transition by activating other repressors of FLC indirectly (Niu et al. 2007; Pei et al. 2007). DRM2 is involved in a conserved epigenetic regulatory mechanism through DNA methylation in Arabidopsis (Zhong et al. 2014).

Fig. 2
figure 2

Through chromatin modification, DRM2, FLD, FVE, HDA5, HDA6, LD, PRMT5, PRMT10 and REF6 can regulate FLC expression. FVE interacts with the histone demethylase FLD and both have been implicated in histone deacetylation complexes, however, the mechanism is unknown recently. The histone deacetylase HDA5 interacts with HDA6 and both display deacetylase activity. HDA6 interacts with FLD. These 4 proteins may present in a protein complex and interplay to repress FLC expression via histone modification including demethylation and deacetylation. REF6 acts as a histone demethylase. PRMT5 functions independently of PRMT10, but both play an important role in asymmetric histone arginine methylation. DRM2 functions in methylation of DNA

Full size image

FVE, also termed as MSI4 (MULTICOPY SUPPRESSOR OF IRA1 4) and ACG1 (ALTERED COLDRESPONSIVE GENE EXPRESSION 1), is orthologous to yeast MSI1 (MULTICOPY SUPPRESSOR OF IRA1) and animal retinoblastoma-associated proteins, which is a component of a histone-acetylation complex. Arabidopsis FVE belongs to a family of MSI1-like WD40 repeat proteins (Kenzior and Folk 1998; Kim et al. 2004; Ausin et al. 2004; Hennig et al. 2005). Together with FLD, FVE may be part of a larger protein complex that represses FLC expression through histone deacetylation (He et al. 2003; Ausin et al. 2004; Amasino 2004). FLD encodes a histone demethylase which is homologous to the human histone H3K4 demethylase LSD1 (LYSINE-SPECIFIC HISTONE DEMETHYLASE1) (Jiang et al. 2007; Yu et al. 2011).

Histone deacetylase HDA5 and HDA6 regulate gene expression cooperately, which display its deacetylase activity by binding to the chromatin of FLC (Yu et al. 2011; Luo et al. 2015). In BiFC (bimolecular fluorescence complementation) and CHIP (Chromatin Immunoprecipitation) assays, HDA5 interacts with HDA6, HDA6 interacts with FLD and FLD interacts with FVE (Yu et al. 2011; Luo et al. 2015). The results indicate that these proteins may present in a co-repressor complex to regulate FLC expression (Fig. 2), suggesting the functional interplay between histone deacetylation and demethylation.

LD is a unique protein possessing a homeodomain-like domain and locates to nucleus in Arabidopsis (Lee et al. 1994; Aukerman et al. 1999). LD was initially considered to be a transcriptional regulator (Blázquezet al. 2001) but was later found to repress FLC expression through a negatively regulatory interaction with SUF4 (SUPPRESSOR OF FRIGIDA 4), a transcriptional activator of FLC (Kim et al. 2006). Until just recently, LD was shown to repress FLC expression via histone modification such as H3 deacetylation and H3K4 demethylation (Domagalska et al. 2007).

Previously,REF6 has been shown to be critical in the regulation of Arabidopsis flowering and act as a FLC repressor. REF6 suppresses FLC transcription through histone modifications in FLC chromatin, suggesting that this class of proteins play the activity of transcriptional regulation by remodeling chromatin (Noh et al. 2004). Recently, REF6 was found to act primarily on FLC antisense RNA (Hornyik et al. 2010a) through its JMJC domain (Jumonji C domain). This domain has been found in human histone demethylases JHDM (jumonji C domain-containing histone demethylase) 2A and JHDM1, and specifically causes H3K9me and H3K36me, respectively (Noh et al. 2004).

Histone H3 lysine methyltransferases are known to be pivotal in gene silencing and developmental control in plants. Recent studies have found that PRMT5 is a type II histone arginine methyltransferase that plays an important role in promoting growth and flowering (Pei et al. 2007). ArabidopsisPRMT10, the Arabidopsis ortholog of plant histone arginine methyltransferase 10 (PHRMT10), a dimeric plant-specific histone H4 methyltransferase in cauliflower; was shown to be a type I PRMT. PRMT10 is found to react in the autonomous pathway and may act as a modulator by activating other repressors of FLC indirectly to control the floral transition. Disruption of PRMT10 resulted in late flowering through the upregulation of FLC transcript levels. Genetically, PRMT10 functions independently of PRMT5, but both act to fine-tune the expression of FLC. This result also indicates the importance of asymmetric arginine methylation in plant development and flowering-time regulation (Niu et al. 2007).

DNA methylation is a classical epigenetic gene regulatory mechanism in the autonomous flowering time pathway. DRM2 (DOMAINSREARRANGEDMETHYLTRANSFERASE2) is a key de novo methyltransferase and is component of a complex possessing the siRNA effector ARGONAUTE4 (AGO4) and preferentially methylating one DNA strand, which likely acts as the template for RNA polymerase V mediated non-coding RNA transcripts. The DNA methylation is positively correlated with the accumulation of strand-biased siRNA. These data indicate that AGO4-siRNA leads DRM2 to its target,and the later in involved in the siRNAs-associated base pairing (Zhong et al. 2014).

Post-translational modification of FLC is mediated by CK2 and PP2A-B’γ

Post-translational modification of proteins has been shown to be indispensable in the regulation of all aspects of plant development including flowering. Casein kinase II (CK2) and Protein phosphatase 2A (PP2A) repress FLC to drive flowering through phosphorylation and dephosphorylation (Mulekar, et al. 2012; Mulekar and Huq 2015).

Protein kinases modify their substrate protein by adding one or more phosphate groups to it, which frequently affects its cellular function and/or abundance. Phosphatases can remove the phosphate group from the substrate proteins. This reversible phosphate group transfer results in post-translational modification of target proteins and allows cells to rapidly respond to aninternal cue and/or external stimulus. Casein kinase II (CK2) is a necessary and highly-conserved Ser/Thr kinase that regulates proteins in the post-translational process in all eukaryotes. Evidence from several prediction algorithms show that the majority of the autonomous pathway components, including FLC, have multiple CK2 phosphorylation sites, which may modulate their activity or stability and thus drive flowering (Mulekar et al. 2012; Mulekar and Huq 2015). Protein phosphatase 2A (PP2A) comprises 3 types of subunits: scaffolding (A), regulatory (B) and catalytic (C) subunit. The knockdown lines of PP2A-b’γ displayed a late flowering phenotype in Arabidopsis. The function of PP2A-B’γ in the autonomous pathway is to repress the main flowering inhibitor, FLC. The knockout lines PP2A-b55α and PP2A-b55β flowered earlier than wild type. These results demonstrate that PP2A acts as both a positive and negative regulator of flowering time, depending on which types of regulatory B subunit is involved (Heidari et al. 2013).

Other regulatory mechanisms in the autonomous flowering pathway

The SYP22 gene encodes a vacuolar N-ethyl-maleimide sensitive factor attachment protein receptor (SNARE) that plays a role in vacuolar and endocytic trafficking pathways. Disruption of SYP22 increases expression of FLC and leads to the late flowering phenotype in Arabidopsis (Ebine et al. 2012). Also the elevated levels of FLC transcripts accumulated in doc1-1 (DARK OVER-EXPRESSION OF CAB 1, DOC1) mutant, and the syp22 phenotypes were enhanced in the syp22 doc1-1 (Ebine et al. 2012). This elevated expression of FLC and the phenotype were suppressed by ara6-1, a mutation in the gene encoding a Rab GTPase involved in endosomal trafficking, indicating the involvment of vacuolar and/or endocytic trafficking in the FLC regulation of flowering (Ebine et al. 2012).

Other 5 genes have also been shown to involve in the autonomous pathway, however the molecular mechanism underlining the regulation need to be elucidated. It is necessary for plant growth and development to be post-translational modified through attachment of a small ubiquitin-like modifier (SUMO) (Villajuana-Bonequi et al. 2014). In Arabidopsis, early genetic analysis indicated that ESD4 (EARLY IN SHORT DAYS 4) is involved in the autonomous pathway. Furthermore, mRNA levels of FLC are decreased by the esd4 mutation, and the expression of flowering-time genes known to be repressed by FLC, are increased in the esd4 mutants (Reeves et al. 2002). Recent research has revealed that ESD4 encodes a SUMO protease, and mutation in this gene causes hyperaccumulation of conjugates formed between SUMO and its substrates. FLC has been shown to interact with the SUMO ligase and to be subsequently modified (Son et al. 2014). Thus, FLC-mediated flowering repression might be positively regulated by sumoylation, mediated by ESD4.

DRM1 (Developmentally Retarded Mutant1) is a novel flowering-promoting locus. The drm1mutation is a single recessive nuclear mutation, and is late flowering under all photoperiod conditions. Moreover, vernalization treatment could restore its late flowering phenotype significantly, suggesting that drm1 is a typical late-flowering mutant and most likely involves in the autonomous flowering pathway. The expression of 3 important repressors, FLC, EMF and TFL1, were increased, in drm1 mutant,.impliying that these repressors act in parallel pathways in the drm1mutant to regulate flowering. It also suggests that DRM1 might be a upstream regulator for these repressors (Zhu et al. 2005).

PRP39-1 (Pre-mRNA Processing Protein 39-1) has been identified and appears to promote flowering indirectly through RNA processing. Mutant lines of PRP39-1 in Arabidopsis show increased expression of FLC accompanied by downregulation of FT and SOC1 (Wang et al. 2007).

Although AGL28 is known to act in the vegetative growth, overexpression of Arabidopsis AGL28 causes early flowering through increasing FCA and LD expression. Hence, AGL28 promotes the autonomous flowering transition (Yoo et al. 2006). However, disruption of AGL28 does not lead to any obvious flowering phenotype, suggesting that AGL28 might be functionally redundant.

DNA-binding protein phosphatase 1 (DBP1) was shown to bind with DNA and displayed protein phosphatase activity in vitro (Carrasco et al. 2006). Zhai et al. (2016) reported that DBP1 was involved in the flowering time regulation via the autonomous pathway and the photoperiod pathway by modifying the transcript levels of several important integrators, such as CO, SOC1, LFY, FT and FLC.

Conclusions and perspective

This review summarizes recent research progress in the autonomous pathway of flowering time regulation in Arabidopsis. Autonomous pathway constituents participate in repressing the main flowering inhibitor FLC and thus, indirectly promote floral transition. Key regulators in this pathway include AGL28, CK2, DBP1,DRM1,DRM2, ESD4, FCA, FLD, FLK, FPA, FVE, FY, HDA5, HDA6, LD, PCFS4, PEP, PP2A-B’γ, PRMT5, PRMT10, PRP39-1, REF6 and SYP22 (Table 1). The molecular mechanisms underlying the regulation of flowering by the autonomous pathway members are primarily concerned with FLC RNA processing mediated by FCA, FLK, FPA, FY, PCFS4 and PEP (Table 1; Fig. 1) and chromatin modification mediated by DRM2, FLD, FVE, HDA5, HDA6, LD, PRMT5, PRMT10 and REF6 (Table 1; Fig. 2); and finally, post-translational modification of FLC mediated by CK2 and PP2A-B’γ (Table 1).

In the future, it can be predicted that other additional members of the autonomous pathway will be identified, and the molecular mechanisms behind previously undefined mediated by AGL28, DBP1, SYP22, ESD4, PRP39-1, DRM1 (Table 1) and newly-discovered members will be revealed. These studies extend our existing understanding of the molecular mechanisms of the autonomous flowering time pathway and may reveal new, as yet undefined, regulatory mechanisms.