Arabidopsis NUCLEAR FACTOR Y A8 inhibits the juvenile-to-adult transition by activating transcription of MIR156s

Abstract

Vegetative (juvenile-to-adult) and flowering (vegetative-to-reproductive) phase changes are crucial in the life cycle of higher plants. MicroRNA156 (miR156) and its target SQUAMOSA PROMOTER BINDING PROTEIN-LIKE (SPL) genes are master regulators that determine vegetative phase changes. The miR156 level gradually declines as a plant ages and its expression is rapidly repressed by sugar. However, the underlying regulatory mechanism of transcriptional regulation of the MIR156 gene remains largely unknown. In this study, we demonstrated that Arabidopsis NUCLEAR FACTOR Y A8 (NF-YA8) binds directly to CCAAT cis-elements in the promoters of multiple MIR156 genes, thus activating their transcription and inhibiting the juvenile-to-adult transition. NF-YA8 was highly expressed in juvenile-stage leaves, and significantly repressed with developmental age and by sugar signals. Our results suggest that NF-YA8 acts as a signaling hub, integrating internal developmental age and sugar signals to regulate the transcription of MIR156s, thus affecting the juvenile-to-adult and flowering transitions.

Flowering, juvenile-to-adult transition, miR156, NF-YA8, phase changes

Introduction

The transitions from the juvenile to the adult stage (also known as the vegetative phase change) and from the adult to the reproductive stage (also known as flowering) are crucial for reproductive success and survival of higher plants (Bäurle and Dean, 2006). In Arabidopsis, the transition from juvenile to adult stages usually involves changes in a variety of morphological and physiological traits, including increases in the complexity of leaf shape (e.g. serrations at leaf margins and trichomes on the leaf surface), and increases in the ratio of leaf length to width (Poethig, 2003; Yu et al., 2015). At adult stages, various environmental signals (e.g. photoperiod, temperature, and stress) and internal cues (e.g. developmental, age, gibberellin, and sugar) coordinately modulate flowering time by regulating transcription of multiple flowering time integrators, such as FLOWERING LOCUS T (FT) and SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1 (SOC1), which further affect the induction of floral primordia at the shoot apical meristem (Mouradov et al., 2002; Imaizumi and Kay, 2006; Amasino, 2010; Song et al., 2015; Cho et al., 2017).

In the vegetative phase change, miR156 and its target SQUAMOSA PROMOTER BINDING PROTEIN-LIKE (SPL) genes are key components that determine the juvenile-to-adult transition and regulate aspects of many other biological processes (Wang et al., 2009; Wu et al., 2009; Xu et al., 2019; Ye et al., 2019). In juvenile-stage leaves, miR156 is present at high levels, thus leading to low levels of SPL mRNA at juvenile stages (Wang et al., 2009; Wu et al., 2009). The abundance of miR156 gradually decreases as plant age increases, thus allowing accumulation of SPL mRNAs at adult stages and the induction of flowering (Wang et al., 2009; Wu et al., 2009; Wang, 2014; Gou et al., 2019). The Arabidopsis genome has eight different loci (MIR156A–H) encoding primary transcripts of miR156 (pri-miR156) (Wu and Poethig, 2006). The juvenile-to-adult transition occurred slightly earlier in mir156a and mir156c single mutants and much earlier in mir156a/c and mir156a/b/c/d multiple mutants (Yu et al., 2013; He et al., 2018). Increasing miR156 abundance or reducing transcript levels of SPLs (such as SPL3, SPL4, SPL5, SPL9, and SPL15) prolongs the juvenile phase and delays flowering (Schwarz et al., 2008; Wang et al., 2009; Wu et al., 2009; Jung et al., 2016). Overexpressing miR156 target mimics (35S::MIM156, producing an mRNA that can specifically bind to miR156) perturbed endogenous miR156 molecular function and significantly increased the abundances of SPL transcripts, thus promoting the juvenile-to-adult transition (Franco-Zorrilla et al., 2007; Wang et al., 2009; Wu et al., 2009; Todesco et al., 2010). Transcription of MIR156s is strictly regulated by environmental and internal signals (Yu et al., 2015; Xie et al., 2017; Xu et al., 2018a; Xu et al., 2018b). Recently, studies revealed that sugar acts as an essential signal repressing MIR156 transcription and promoting the juvenile-to-adult transition (Yang et al., 2013; Yu et al., 2013, 2015; Buendía-Monreal and Gillmor, 2017); however, the underlying regulatory mechanism remains largely unknown.

The heterotrimeric transcription factor Nuclear Factor Y (NF-Y, also known as CCAAT-box binding factor, CBF) is composed of NF-YA, NF-YB, and NF-YC subunits and is evolutionarily conserved in all higher eukaryotes (Edwards et al., 1998; Petroni et al., 2012; Laloum et al., 2013). Yeast and animal genomes generally have only one or two genes encoding each NF-Y subunit (Li et al., 2018a). By contrast, flowering plant genomes have multiple genes encoding each NF-Y subunit (Petroni et al., 2012; Laloum et al., 2013). For example, the Arabidopsis genome has 10 distinct genes encoding each NF-Y subunit; these different subunits could potentially form a thousand diverse NF-YA/B/C trimeric complexes and regulate numerous cellular processes (Edwards et al., 1998; Hackenberg et al., 2012; Petroni et al., 2012; Laloum et al., 2013). Although NF-Y transcription factors have been identified in most species and function in diverse biological processes, their complicated regulatory mechanisms remain mysterious. Previous genetic and molecular studies have revealed that each subunit of NF-Y may exhibit specific physiological functions by affecting transcription of various downstream targets, and these functions cannot be entirely replaced by other members of the same subunit family (Laloum et al., 2013; Zhao et al., 2016). In flowering time regulation, Arabidopsis NF-YB1 inhibits flowering by repressing FT transcription (Wenkel et al., 2006), while B2 and B3 promote flowering by activating FT transcription (Kumimoto et al., 2008). Interestingly, B9/LEC1 and B6/LEC1-like interact with other B subunits (B1, B8, B10) to repress flowering mainly by activating FLOWERING LOCUS C (FLC) transcription (Tao et al., 2017). Arabidopsis NF-YC1, C2, C3, C4, and C9 promote photoperiod-dependent flowering by activating the transcription of FT and SOC1 (Kumimoto et al., 2010; Gnesutta et al., 2017; Hwang et al., 2019), but overexpressing NF-YC10 did not produce significant changes in flowering time (Sato et al., 2014). In addition, overexpressing Arabidopsis NF-YA1, A3, A4, A5, A7, A9, and A10, delayed flowering time, indicating that the NF-YA subunits they encode negatively regulate flowering time (Wenkel et al., 2006; Leyva-González et al., 2012; Mu et al., 2013; Xu et al., 2014), while the effects of A2, A6, and A8 subunits on flowering remain unclear or contradictory, based on several previous studies (Mu et al., 2013; Hou et al., 2014; Xu et al., 2014; Siriwardana et al., 2016). Notably, no obvious early flowering phenotype has been reported for single mutants of these NF-YA subunits (Siefers et al., 2009; Fornari et al., 2013; Mu et al., 2013). Although most NF-YA subunits play crucial roles in flowering time regulation, the underlying regulatory mechanisms remain largely unknown.

In this study, we found that the Arabidopsis NF-YA8 subunit negatively regulated flowering time and repressed the juvenile-to-adult transition. We demonstrated that NF-YA8 directly bound to the CCAAT cis-element in the promoters of MIR156A, C, D, and E, activating their expression and thereby prolonging the juvenile phase. Interestingly, the transcript level of NF-YA8 gradually declined as the plant age increased, and its expression was rapidly repressed by increased sugar. These results suggest that Arabidopsis NF-YA8 may act as a hub integrating developmental age and sugar signals to regulate the juvenile-to-adult transition and flowering time.

Materials and methods

Plant materials and growth conditions

Arabidopsis ecotypes Columbia (Col-0) and Nossen (No-0) were used in this study. Transgenic plants 35S::NF-YA8-3MYC, NF-A8p::YA8-LUC, NF-YA8p::GUS, and MIR156Ap::GUS were generated in the No-0 ecotype. Arabidopsis T-DNA insertion mutants nf-ya2 (SALK_021228C), nf-ya3 (SALK_068206), nf-ya8-1 (CS876601), nf-ya8-2 (CS873143), nf-ya10 (SALK_127699C), and 35S::MIM156 transgenic plants were in the Col-0 ecotype background and have been described previously (Franco-Zorrilla et al., 2007; Fornari et al., 2013). Arabidopsis seeds were sterilized and sown on Murashige and Skoog (MS) solid medium containing 1% sucrose after stratification at 4 °C overnight, and then transferred to long-day (LD, 16 h light–8 h dark) or short-day (SD, 8 h light–16 h dark) growth conditions at 22 °C. For juvenile-to-adult transition and flowering time measurement, Arabidopsis seeds were directly planted in soil after stratification at 4 °C, then germinated and grown under either LD or SD conditions. The juvenile-to-adult transition was measured based on the presence of a serrated margin and abaxial trichomes on emerging leaves. Flowering time was measured by counting the number of rosette leaves and days from germination to the first floral bud opening.

Plasmid construction and transformation

To generate 35S::NF-YA8-3MYC, the full-length coding region of NF-YA8 was PCR-amplified from cDNA of Arabidopsis No-0 plants and then ligated into KpnI- and BcuI-digested pCAMBIA1305-3MYC (Li et al., 2018b) to produce the 35S::NF-YA8-3MYC binary vector. To generate NF-YA8p::GUS, a 2-kb fragment upstream of the transcription start site of NF-YA8 was PCR-amplified from genomic DNA of No-0 plants and then ligated into KpnI- and BamHI-digested pPZP211-GUS (Ma et al., 2016) to produce the NF-YA8p::GUS binary vector. To generate MIR156Ap::GUS, a 1.8-kb fragment upstream of the transcription start site of MIR156A was PCR-amplified from genomic DNA of No-0 plants and then ligated into BamHI- and SalI-digested pPZP211-GUS to produce the MIR156Ap::GUS binary vector. To generate NF-YA8p::YA8-LUC, a 2-kb fragment of the NF-YA8 promoter was first ligated into KpnI- and BamHI-digested pPZP211-LUC (Chen et al., 2019) to produce pPZP211-A8p::LUC. Then, the full-length coding region of NF-YA8 was PCR amplified from 35S::NF-YA8-3MYC and ligated into BamHI- and Bcul-digested pPZP211-YA8p::LUC to produce the pPZP211-NF-A8p::YA8-LUC binary vector. All these various binary vectors were transformed into Arabidopsis No-0 wild type (WT) plants by the Agrobacterium-mediated floral-dip method. The primers used in the study are shown in Supplementary Table S1 at JXB online.

RNA-seq and data analysis

For RNA-seq assays, 10-day-old No-0 and NF-YA8-3MYC (A8-O1) transgenic plants grown under LD conditions were harvested at zeitgeber time 10 (ZT10). RNA extraction and cDNA library generation were performed according to procedures described previously (Kong et al., 2017). The Illumina HiSeq 2500 system was used for RNA-seq at Annoroad Gene (Beijing). Analysis of RNA-seq data was performed following the procedure described previously (Kong et al., 2017). Briefly, all the reads were used for mapping to the Arabidopsis genome (TAIR10) with HISAT2 and then unique reads with a perfect match or one mismatch were further analysed and annotated. DESeq2 was employed for differentially expressed gene analyses with fold-change >2 and false discovery rate <0.05.

RNA extraction and RT-qPCR

Total RNA was isolated using a plant RNA extraction kit (CWBIO), and the first-strand cDNA was then synthesized using an EasyScript One-Step gDNA Removal and cDNA Synthesis SuperMix kit (Transgene) following the manufacturer’s instructions. RT-qPCR assays were further conducted using UltraSYBR Mixture (CWBIO) in a LightCycler 96 PCR machine (Roche). Gene expression values were calculated as described previously (Tian et al., 2020). UBQ1 was used as an internal control for RT-qPCR assays.

Chromatin immunoprecipitation–qPCR

Seven-day-old NF-YA8-3MYC plants grown under LD conditions were used to perform chromatin immunoprecipitation (ChIP)–qPCR assays. ChIP was conducted according to a previously described procedure (Li et al., 2011) using monoclonal anti-MYC antibodies (Cell Signaling Technology). A ChIP reaction with no antibodies was used as a negative control for the ChIP procedure. The qPCR analyses of ChIP products were conducted using TransStart Tip Green qPCR SuperMix (Transgene) in a LightCycler 96 PCR machine (Roche). UBQ1 was used as a negative control for the ChIP-qPCR assay.

β-Glucuronidase activity assay

Seven-day-old seedlings of NF-YA8p::GUS or MIR156Ap::GUS reporter lines were used for GUS staining or quantification assays. These seedlings were harvested at the indicated times after 100 mM glucose (Glc) or 100 mM mannitol (Man) treatment and then incubated in β-glucuronidase (GUS) staining solution at 37 °C for 8 h. After GUS staining, various plants were decolorized in 70% ethanol and then photographed. The activity of GUS was measured following a procedure described previously (Wu and Poethig, 2006).

Transient dual-luciferase expression assay

To generate LUC reporters for the transient assays, various fragments containing the promoter regions of MIR156A, MIR156C, and SPL3, 5, 9, and 15 were amplified from genomic DNA of No-0 and then ligated into KpnI- and SalI-digested pGreenII 0800-LUC (Hellens et al., 2005). Various fragments containing WT or mutated CCAAT cis-elements in the promoter regions of MIR156 were synthesized by Sangon Biotech (Shanghai) and then ligated into KpnI- and SalI-digested pGreen-0800-35Smini vector (Chen et al., 2019).

For transient expression assays in Arabidopsis protoplasts, 21-day-old WT No-0 plants grown under SD conditions were used to prepare protoplasts following a procedure described previously (Hellens et al., 2005). The plasmids of various LUC reporters (10 μg) and effectors (15 μg 35S::NF-YA8-MYC or MYC) were co-transformed into Arabidopsis protoplasts. After incubation for 16 h at 22 °C, the bioluminescence intensity of LUC was determined with the Promega Dual-Luciferase Report Assay system and a GloMax 20/20 luminometer. For the transient expression assay in Nicotiana benthamiana leaves, the various effector and reporter constructs were transformed into Agrobacterium strain GV3101 and then co-infiltrated into the leaves of N. benthamiana. After infiltration for 2 d, the bioluminescence intensity of LUC was detected by Cool-CCD camera (NightSHADE LB985). Five independent biological replicates were carried out for each combination in transient expression assays.

Accession numbers

The Arabidopsis Genome Initiative locus identifiers for the genes mentioned in this article are as follows: NF-YA8, AT1G17590; MIR156A, At2G25095; MIR156C, AT4G31877; MIR156D, AT5G10549; MIR156E, AT5G11977; SPL3, AT2G33810; SPL4, AT1G53160; SPL5, AT3G15270; SPL9, AT2G42200; SPL15, AT3G57920; UBQ1, AT3G52590.

Statistical analysis

For comparing the difference between two groups, Student’s t-test was used to perform statistical analysis. Statistically significant differences are indicated with one asterisk at P<0.05 and two asterisks at P<0.01. When determining the difference among multiple groups, one-way ANOVA with Duncan’s multiple range test (P<0.05) was used, and statistical differences were indicated with distinct letters.

Results

Arabidopsis NF-YA8 negatively regulates plant growth and flowering time

We started our investigation of the physiological functions of Arabidopsis NF-YA8 with two T-DNA insertion mutants, nf-ya8-1 (CS876601) and nf-ya8-2 (CS873143). After confirming the absence of NF-YA8 gene expression in the a8-1 and a8-2 lines (see Supplementary Fig. S1A), our phenotypic analysis revealed that neither of the mutants exhibited significant changes in flowering time compared with Col-0 plants under either LD or SD conditions (Supplementary Fig. S1B). Further, we obtained T-DNA insertion mutants nf-ya2 (SALK_021228C), nf-ya3 (SALK_068206), and nf-ya10 (SALK_127699C) from the Arabidopsis Biological Resource Center (ABRC), and genetically crossed a2 with a8-2 and a10 mutants. We did not notice any significant changes in flowering time whether in a2, a3, and a10 single mutants or in a2 a10 and a2 a8 double mutants compared with WT Col-0 plants under either LD or SD growth conditions (data not shown). Given that multiple NF-YA subunits share high amino acid identities (Siefers et al., 2009; Petroni et al., 2012; Laloum et al., 2013), it is likely that high functional redundancy exists among these A-subunits of NF-Y, which could mask a particular role of NF-YA8.

To further investigate the physiological function of Arabidopsis NF-YA8, we generated NF-YA8-3MYC overexpression lines in the No-0 ecotype (see Supplementary Fig. S1C, D). Compared with No-0 control plants, NF-YA8-3MYC lines (A8-O1, A8-O7, and A8-O8) exhibited obvious retardation of growth, such as reduced leaf size and petiole length, and altered shape of leaf blades (Fig. 1A, B). Moreover, NF-YA8-3MYC transgenic plants dramatically delayed flowering under LD and SD growth conditions, indicating that NF-YA8 negatively regulates plant growth and flowering time in Arabidopsis (Fig. 1C, D).

Fig. 1.

Arabidopsis NF-YA8 negatively regulates plant growth and flowering time. (A) Photographs of whole plants and representative leaves (L7, L9, and L11) of NF-YA8-3MYC plants. Scale bar: 1 cm. (B) Measurements of leaf area and petiole length (L7) of NF-YA8-3MYC plants. O1, O7, and O8 are three independent transgenic lines of 35Sp::NF-YA8-3MYC. No-0, wild type (WT) control plant. Values are means ±SD, n=20. (C, D) Photographs (C) and measurements of rosette leaf number and days to flowering (D) showing delayed flowering phenotype of NF-YA8-3MYC plants compared with No-0 plants. Values are means ±SD; n=20. Significant different are indicated with distinct letters in (B, D) (one-way ANOVA, Duncan’s multiple range test, P<0.05).

Genome-wide analysis of NF-YA8 downstream targets in flowering time regulation

To investigate the molecular mechanism by which NF-YA8 negatively regulates growth and flowering time, we performed RNA-seq using 10-day-old NF-YA8-3MYC transgenic plants (A8-O1) and WT No-0 plants grown under LD conditions. There were 22 314 and 22 364 expressed genes detected in NF-YA8-3MYC and WT plants, respectively (see Supplementary Table S2). Based on fold-change >2 and false discovery rate <0.05, 1232 up-regulated, and 1247 down-regulated genes were identified in NF-YA8-3MYC plants compared with WT No-0 plants (Supplementary Table S3). Notably, the expression levels of a large proportion of flowering time-related genes, including FT, SOC1, and FLC, were obviously altered in NF-YA8-3MYC plants compared with WT No-0 plants (Supplementary Table S4; Fig. 2A). Further, the expression profiles of multiple flowering integrators and various flowering pathway genes were verified in NF-YA8-3MYC plants using RT-qPCR (Fig. 2B). Both RNA-seq and RT-qPCR results showed that the transcript levels of flowering promoting genes FT and SOC1 were significantly decreased, while the transcript level of flowering repressing gene FLC was significantly increased, in NF-YA8-3MYC plants compared with No-0 control plants (Fig. 2A, B). However, the transcript levels of most vernalization (e.g. FRI), photoperiod (e.g. CO), gibberellic acid (e.g. RGA), and autonomous (e.g. FY) flowering pathway genes were not obvious altered in NF-YA8-3MYC plants compared with the No-0 control plants (Fig. 2A, B; Supplementary Table S5). Interestingly, a large proportion of age-dependent genes, including SPL2, SPL3, SPL4, SPL5, SPL8, SPL9, SPL11, and SPL15, showed lower transcript levels in NF-YA8-3MYC plants—only about 23–50% of the levels observed in No-0 control plants (Fig. 2A; Supplementary Tables S4, S5). These results suggested that NF-YA8 negatively regulates flowering time and this might be by affecting the expression of these flowering integrators and age-dependent flowering pathway genes.

Fig. 2.

Expression of flowering time-related genes in NF-YA8-3MYC plants. (A) Heatmap showing the expression profiles of selected flowering time-related genes in NF-YA8-3MYC (A8-O1) plants, compared with No-0 plants. The key below shows down-regulation to the left and up-regulation to the right. R1, R2, and R3 represent three replicates of RNA-seq assays. (B) RT-qPCR analyses showing the relative expression of flowering time-related genes in NF-YA8-3MYC (A8-O1 and A8-O7) plants. Values are means ±SD of three technical replicates. Significant differences of each detected gene are indicated with distinct letters (one-way ANOVA, Duncan’s multiple range test, P<0.05). This experiment was independently repeated three times. Ten-day-old LD grown plants were used to perform RNA-seq (A) and RT-qPCR (B) assays.

NF-YA8 delays the transition from juvenile to adult stage

To verify whether NF-YA8 regulates flowering time through an age-dependent pathway, we first investigated the transition from juvenile to adult stages in NF-YA8-3MYC plants by comparing the leaf morphology of NF-YA8-3MYC plants with WT No-0 control plants. In Arabidopsis, rosette leaves in the adult stage are elongated with a serrated margin and trichomes on the abaxial surface (Wu and Poethig, 2006; Guo et al., 2017). After 21 d of growth under LD conditions, 100% of the first six rosette leaves and 89% of the seventh rosette leaf of WT No-0 plants had no clear marginal serration or abaxial trichomes and were in the juvenile stage. By contrast, in NF-YA8-3MYC plants, 100% of the first eight rosette leaves and about 70% of the ninth leaf were in the juvenile stage, indicating that overexpression of NF-YA8 postponed the juvenile-to-adult transition (Fig. 3A, B; Supplementary Fig. S2).

Fig. 3.

Arabidopsis NF-YA8 inhibits the transition from the juvenile stage to the adult stage. (A, B) Morphology of rosette leaves (A, left, white arrowheads indicate serrated margins), presence of abaxial trichomes (A, right, in black), and percentage of juvenile-stage leaves (B) in 3-week-old No-0 and NF-YA8-3MYC (A8-O1, A8-O7, and A8-O8) plants grown under LD. Values are means ±SD (n=10). (C) RT-qPCR analyses showing the transcript abundances of SPL3, SPL9, MIR156A, and MIR156C in the rosette leaves (leaf 5–17) of 3-week-old No-0 and NF-YA8-3MYC plants grown under LD. (D) RT-qPCR analyses showing the transcript abundances of SPLs (left) and MIR156s (right) in 7-day-old LD plants of NF-YA8-3MYC (A8-O1 and A8-O7) compared with control No-0 plants. In (C, D), values are means ±SD of three technical replicates. Significant differences of each detected gene are indicated with distinct letters in (D) (one-way ANOVA, Duncan’s test, P<0.05). These experiments were independently repeated three times.

miR156 and its target SPLs play crucial roles in the age-dependent flowering pathway by affecting the transition from juvenile to adult stages (Wang et al., 2009; Wu et al., 2009). To investigate whether NF-YA8 represses the juvenile-to-adult transition by affecting the miR156–SPLs regulatory module, we used RT-qPCR to measure the transcript levels of MIR156A, MIR156C, SPL3, and SPL9. In NF-YA8-3MYC plants, the transcript abundances of SPL3 and SPL9 were lower compared with those of WT plants, consistent with its delayed juvenile-to-adult transition and late flowering (Fig. 3C). The transcript levels of MIR156A and MIR156C were dramatically higher in NF-YA8-3MYC plants compared with No-0 control plants, indicating that decreased accumulation of SPL3 and SPL9 was caused by increased transcription of MIR156s (Fig. 3C). RT-qPCR analyses also revealed that the transcript levels of SPL4, SPL5, and SPL15 were significantly lower in NF-YA8-3MYC plants compared with No-0 control plants (Fig. 3D), consistent with the RNA-seq results (see Supplementary Table S4). Interestingly, the transcript levels of MIR156A, C, D, and E, but not MIR156B, F, and G, were obvious higher in NF-YA8-3MYC plants, indicating that NF-YA8 positively regulates their transcription (Fig. 3D). These results suggest NF-YA8 delays flowering by repressing the juvenile-to-adult transition, and this is likely to be achieved through effects on the miR156–SPLs regulatory module.

NF-YA8 directly binds to MIR156 promoters and activates their transcription

NF-YA8 and miR156 repress the juvenile-to-adult transition, and overexpression of NF-YA8 increases the expression of MIR156s (Fig. 3C, D); these observations suggest that NF-YA8 may activate the transcription of MIR156s directly. Promoter sequence analyses revealed that multiple NF-Y transcription factor binding cis-elements (CCAAT) were identified within 2 kb upstream of the transcription start site for MIR156A, C, D, and E (Fig. 4A). Although a DNA binding domain exists at the C-terminus of NF-YA subunits, yeast one-hybrid assays failed to verify direct binding of NF-YA8 to the promoters of MIR156s (data not shown), possibly because NF-Y function as a trimeric transcription factor requires all three of the NF-YA, B, and C subunits (Hackenberg et al., 2012). We further performed ChIP assays that verified that the NF-YA8 subunit directly binds to fragments containing the CCAAT cis-elements in the promoter regions of MIR156A (A3), C (C1, C2, and C3), D (D1 and D2), and E (E2 and E3) (Fig. 4B; Supplementary Fig. S3).

Fig. 4.

NF-YA8 directly binds to the promoters of MIR156 genes and activates their transcription. (A) Schematic diagrams showing the promoter regions of MIR156A, C, D, and E. Vertical lines indicate NF-Y binding cis-elements (CCAAT); solid and dashed lines indicate the fragments used in ChIP-qPCR (B) and transcriptional activation assays (C), respectively. (B) ChIP-qPCR assays showing the enrichment of specific fragments of MIR156s promoters by immunoprecipitation of NF-YA8–3MYC. Anti-MYC antibodies were used to perform ChIP in 7-day-old NF-YA8-3MYC plants grown under LD conditions. UBQ1 was used as a negative control. Values are means ±SD of three technical replicates. **P<0.01. This experiment was repeated three times. (C) Transcriptional activation assays in Arabidopsis protoplasts showing NF-YA8 directly activates the expression of LUC reporters driven by various fragments of MIR156s promoters. Am, C4m, C5m, D4m, and E5m indicate fragments containing mutated CCAAT cis-elements. MYC indicates the empty vector of A8-MYC used as a negative control for transient expression assays. All fragments of MIR156 promoters are shown in (A). The expression values of REN reporter driven by 35S promoter were used as an internal control. Values are means ±SD (n=4). *P<0.05; **P<0.01.

Next, we performed transit transcriptional activation assays using dual-luciferase (LUC/REN) reporters to test whether NF-YA8 could directly activate the transcription of MIR156 in Arabidopsis protoplasts (Fig. 4C). Various fragments from the promoter regions of MIR156A, C, D, and E containing WT or mutated CCAAT cis-elements were used to drive a LUC reporter. Co-transformation of 35S::NF-YA8-3MYC effector with various MIR156p::LUC reporters resulted in significant activation of the LUC reporter (Fig. 4C). Mutated CCAAT cis-elements in MIR156 promoters failed to activate expression of LUC, demonstrating that these cis-elements are essential for the transcriptional activation of MIR156 by NF-YA8 (Fig. 4C, right). We also verified that NF-YA8 can activate the expression of miR156A, C, D, and E in N. benthamiana leaves (see Supplementary Fig. S4A, B). In addition, transcriptional activation assays showed that NF-YA8 did not regulate the transcription of SPLs directly (Supplementary Fig. S4C). All these results indicate that NF-YA8 inhibits the juvenile-to-adult transition by directly binding to the promoters of MIR156 and activating their transcription.

NF-YA8 delays the juvenile-to-adult transition through MIR156s

To test whether NF-YA8 delays the juvenile-to-adult transition by activating the transcription of MIR156s, NF-YA8NF-YA8-3MYC (A8-O1) plants were genetically crossed with MIM156 plants, in which a target-site mimic of miR156 is overexpressed (Franco-Zorrilla et al., 2007). Under LD conditions, MIM156 transgenic plants exhibited an accelerated juvenile-to-adult transition and early flowering phenotype (Fig. 5A, B), consistent with previous studies (Franco-Zorrilla et al., 2007; Wang et al., 2009; Wu et al., 2009). A8-O1 plants showed a delayed juvenile-to-adult transition and late flowering phenotype with nearly 20 rosette leaves at bolting, whereas A8-O1 MIM156 plants exhibited an earlier juvenile-to-adult transition and early flowering phenotype, similar to MIM156 plants (Fig. 5A, B). In addition, both MIM156 and A8-O1 MIM156 plants produced adult-stage leaves with serrated margins and abaxial trichomes after leaf 2 or 3, distinct from A8-O1 and WT plants (Fig. 5C). The percentage of juvenile-stage leaves was significantly lower in A8-O1 MIM156 plants and similar to MIM156 plants (Fig. 5D).

Fig. 5.

NF-YA8 represses the juvenile-to-adult transition through MIR156s. (A) Morphology of 3-week-old Col-0, MIM156, NF-YA8-3MYC (A8-O1), and A8-O1 MIM156 plants grown under LD conditions. (B) Number of rosette leaves of various plants. Data are means ±SD; n=20. (C, D) Morphology (C, left, white arrowheads indicate serrated margins), presence of abaxial trichomes (C, right, in black), and percentages of juvenile leaves (D) in 3-week-old plants grown under LD conditions. Data in (D) are means ±SD (n=10). (E) Morphology of 18-day-old plants of WT Col-0, MIM156 and NF-YA8-3MYC/MIM156 (1^#–6^#) transgenic lines grown under LD conditions. White arrowheads indicate serrated margins. (F) RT-qPCR analyses showing the relative expression of NF-YA8, MIR156A, and SPL3 in MIM156 and NF-YA8-3MYC/MIM156 lines in 18-day-old plants grown under LD conditions. Values are means ±SD of three technical replicates. Significant differences are indicated with distinct letters in graphs (B, D, F) (one-way ANOVA, Duncan’s multiple range test, P<0.05).

To exclude the effect of ecotype differences between A8-O1 (No-0) and MIM156 (Col-0) plants, we further constitutively expressed NF-YA8 in the MIM156 background (Fig. 5E). Although the transcription of NF-YA8 was obviously increased in the MIM156 background (Fig. 5F), these transgenic plants did not show a delayed juvenile-to-adult transition and late flowering phenotype (Fig. 5E), compared with NF-YA8-3MYC transgenic plants in the WT background (Fig. 1C, D). Rather, the transcription of MIR156A was still constitutively activated by NF-YA8 in the NF-YA8-3MYC/MIM156 plants, but its function might be blocked by MIM156, thus resulting in accumulation of SPL3 and an early transition from juvenile to adult stage (Fig. 5E, F). These results revealed that miR156 acts downstream of NF-YA8 and mediates its function in the juvenile-to-adult transition.

Transcription of NF-YA8 is regulated by developmental age and sugar signals

To investigate the spatial and temporal regulatory mechanisms of the NF-YA8–MIR156s transcriptional cascade, we first measured the expression of NF-YA8 and MIR156A in different rosette leaves. RT-qPCR analyses revealed that both NF-YA8 and MIR156A were expressed the highest in the juvenile stage leaf (fifth leaf) and dramatically decreased with increasing leaf age, being lowest in the adult stage leaf (12th leaf) (Fig. 6A). To further investigate the spatial and temporal expression of NF-YA8 and MIR156s, we generated NF-YA8p::GUS and MIR156Ap::GUS transgenic plants. GUS activity analyses revealed that both NF-YA8 and MIR156A were highly expressed in the vascular tissues of cotyledons and juvenile-stage leaves (L4), and poorly expressed in adult-stage leaves (L9) (Fig. 6B). To further verify whether NF-YA8 protein accumulated substantially in juvenile-stage leaves, we generated NF-YA8p::YA8-LUC transgenic plants. Under LD growth conditions, accumulation of the NF-YA8–LUC fusion protein was considerable in juvenile-stage leaves but drastically decreased with increasing development age, consistent with the expression pattern of NF-YA8 transcript (Fig. 6C). These results revealed that NF-YA8 and MIR156 are expressed in the same tissues and at the same stages, and NF-YA8 mediates the developmental age signal to regulate the transcription of MIR156.

Fig. 6.

NF-YA8 integrates age and sugar signals in Arabidopsis. (A) RT-qPCR analyses showing the relative expression of NF-YA8 and MIR156A in rosette leaves (L5–L12) of 3-week-old plants grown under LD. Significant differences are indicated with distinct letters (one-way ANOVA, Duncan’s multiple range test, P<0.05). (B) GUS staining showing NF-YA8p::GUS and MIR156Ap::GUS expression in cotyledons and juvenile-stage leaves (L4 and L9). (C) Luminescence activity analysis showing the abundance and distribution of NF-YA8p::YA8-LUC in rosette leaves. (D) RT-qPCR analyses showing the transcript abundances of MIR156A, MIR156C, and NF-YA8 after glucose (Glc) or mannitol (Man) treatment in WT plants. (E) GUS staining (left) and activity (right) analyses showing the activity of NF-YA8p::GUS after Glc or Man treatment. In (D, E), 7-day-old LD grown various plants were treated with 100 mM Glc or 100 mM Man for 24 h. Values are means ±SD of three technical replicates, **P<0.01. These experiments were independently repeated three times.

Previous studies demonstrated that sugar signals promote the juvenile-to-adult transition by repressing the transcription of MIR156 in Arabidopsis (Yang et al., 2013; Yu et al., 2013). To explore whether the transcription of NF-YA8 is repressed by sugar signals, 7-day-old WT plants were treated with 100 mM glucose (Glc) or 100 mM mannitol (Man) for 24 h. RT-qPCR analyses revealed that transcript levels of MIR156A, MIR156C, and NF-YA8 were strongly repressed by Glc (Fig. 6D). Glc treatment also significantly reduced the expression of NF-YA8p::GUS in the vascular tissue of cotyledons compared with Man treatment (Fig. 6E). These results revealed that the transcript levels of NF-YA8 and MIR156 are regulated by internal developmental age and sugar signals.

Discussion

Higher plants undergo a series of phase transitions during post-embryonic development; these are critical for reproductive success (Bäurle and Dean, 2006; Wu et al., 2009). Increasing evidences suggest that internal developmental age and sugar signals play essential roles in the juvenile-to-adult transition; however, the underlying molecular mechanisms are still poorly understood. In this study, we demonstrated that Arabidopsis NF-YA8 integrates age and sugar signals, negatively regulating vegetative phase changes and flowering. We found that NF-YA8 bound directly to CCAAT cis-elements in the promoters of MIR156 (A, C, D, and E), activating their transcription and thus inhibiting the juvenile-to-adult transition and delaying flowering time. We also showed that the expression of NF-YA8 was significantly repressed by internal developmental age and sugar signals, suggesting that NF-YA8 may serve as a signal hub integrating age and sugar signals to regulate the expression of MIR156s, thus affecting the juvenile-to-adult transition.

NF-YA8 negatively regulates the juvenile-to-adult transition and flowering

In the cytosol, NF-YB subunits interact with NF-YC subunits through histone-fold domains forming heterodimers. Further, these heterodimers translocate into the nucleus, where they interact with NF-YA subunits forming a trimeric transcription factor that binds to DNA and regulates the transcription of various downstream targets (Myers and Holt, 2018). Although NF-YA, YB, and YC subunits had been supposed to share similar functions, increasing genetic evidence suggests that the various subunits have diverse functions and even act in opposition to each other (Zhao et al., 2016; Myers and Holt, 2018). Moreover, in animal systems, there are one or two genes encoding each subunit and their physiological roles have been extensively studied (Li et al., 2018a). In plants, the numbers of genes encoding each subunit of NF-Y has expanded; therefore, the functions of each subunit are likely to be more complicated and difficult to decipher in plants than in animals (Myers and Holt, 2018). In flowering time regulation, most NF-YA subunits play a negative role (Wenkel et al., 2006; Leyva-González et al., 2012; Mu et al., 2013), most NF-YC subunits play a positive role (Kumimoto et al., 2010; Hou et al., 2014; Gnesutta et al., 2017), and various NF-YB subunits play either a positive (B2 and B3) or a negative (B1, B6, B8, B9, and B10) regulatory role (Wenkel et al., 2006; Kumimoto et al., 2008; Tao et al., 2017). These observations suggest that different NF-Y subunits may function in diverse trimeric complexes or pathways to regulate flowering time.

Here, we revealed that Arabidopsis NF-YA8 negatively regulates flowering time, although a previous study showed that constitutive expression of NF-YA8 had no effect on flowering time regulation (Siriwardana et al., 2016). Under both LD and SD conditions, overexpression of NF-YA8 caused an obvious delayed flowering phenotype, which suggested that it represses flowering time in a photoperiod-independent manner (Fig. 1). Consistent with this notion, most photoperiod-related genes including GIGANTEA (GI), CYCLING DOF FACTOR 1 (CDF1), and EARLY FLOWERING 3 (ELF3) were not obviously regulated by NF-YA8 (Fig. 2, Supplementary Table S5). Although overexpression of NF-YA8 obvious delayed flowering time, the flowering time of its T-DNA insertion mutants is indistinguishable from WT control plants (see Supplementary Fig. S1B), which is possibly caused by functional redundancy among different NF-YA subunits. NF-YA3 is the closest homolog to NF-YA8; the single T-DNA insertion mutants a3 and a8 show no visible defects, while a3 a8 double mutant are embryo-lethal (Fornari et al., 2013). In this study, we generated the double mutants a2 a10 and a2 a8; no obvious altered flowering time phenotype was observed for these plants under either LD or SD growth conditions (data not shown), which suggested that higher order mutants may be required to show the native regulatory role of NF-YA subunit in flowering time regulation. The transcripts of most NF-YA subunits, including NF-YA1, A2, A3, A5, A8, A9, and A10, are direct targets of miR169, and overexpression of miR169 decreases the abundance of these NF-YA transcripts, causing an early flowering phenotype (Zhao et al., 2011; Sorin et al., 2014; Xu et al., 2014). This indicates that a reduction in the transcript levels of most of the A-subunits can lead to an early flowering phenotype.

Interestingly, our study revealed that NF-YA8 negatively regulates flowering time by directly activating MIR156 transcription, thus repressing the juvenile-to-adult transition (Figs 3–5). Overexpression of NF-YA8 delayed the juvenile-to-adult transition and inhibited the production of rosette leaves with serrated margins and abaxial trichomes (Fig. 3). Further, we demonstrated that NF-YA8 directly binds to MIR156 promoters through CCAAT cis-elements and activates their transcription, thus negatively regulating the juvenile-to-adult transition (Fig. 4). Consistent with this, the transcript levels of multiple MIR156s (MIR156A, C, D, and E) were significantly higher than WT in NF-YA8-3MYC plants (Fig. 3). In addition, genetic evidence showed that disruption of miR156 activity (by expression of a target mimic transcript) impaired the negative roles of NF-YA8 in the juvenile-to-adult transition and flowering time regulation (Fig. 5). This further verified that MIR156s is the downstream component mediating NF-YA8 function in these phase changes. Therefore, we conclude that NF-YA8 directly binds to MIR156 promoters to activate their transcription thus mediating juvenile-to-adult transition and flowering time regulation in the age-dependent pathway.

A previous study revealed that an NF-YB8 subunit negatively regulated flowering time by activating transcription of MIR156s in Chrysanthemum morifolium (Wei et al., 2017). The regulatory mechanism by which NF-Y represses phase changes by activating MIR156 transcription therefore exists in both Arabidopsis andChrysanthemum morifolium, two distinct flowering plants (Arabidopsis is a LD flowering plant and Chrysanthemum is a SD flowering plant). Interestingly, promoter sequence analyses of MIR156A in Zea mays, Brachypodium distachyon, Cucumis melo, and Solanum tuberosum showed that multiple CCAAT cis-elements existed in their promoter regions (see Supplementary Fig. S5). This suggests that the transcriptional module of NF-Y–MIR156 may be evolutionarily conserved for phase changes in flowering plants.

In this study, we also noticed that the transcript level of FLC was significantly higher in NF-YA8 overexpression plants (Fig. 2), consistent with a previous study showing that NF-YA subunits negatively regulate flowering time possibly by activating transcription of FLC under stress conditions (Xu et al., 2014). This suggested that FLC might be one of the downstream targets of NF-YA8 mediating flowering time regulation. The transcription of FLC can be activated by NF-YB1, B6, B8, B9, and B10 (Tao et al., 2017), which suggested that NF-YA8 might interact with these NF-YB subunits and function in the same or similar trimeric complex to regulate the transcription of FLC. Indeed, a previous study revealed that NF-YA8 may interact with NF-YB6/L1L and NF-YB9/LEC1 (Sato et al., 2014), consistent with this hypothesis. In addition, previous studies have revealed that Arabidopsis NF-YA8 and its closest homolog, NF-YA3, share similar functions with NF-YB9/LEC1 and its closest homolog, NF-YB6/L1L, in embryogenesis, since both a3 a8 and b6 b9 are lethal at embryo stage (Fornari et al., 2013; Pelletier et al., 2017). At seedling stage, induction of the transcription of NF-YB9/LEC1 repressed the production of trichomes (Junker et al., 2012), consistent with the role of NF-YA8 in the juvenile-to-adult transition (Fig. 3). All this suggests that NF-YA8 might interact with NF-YB6 or B9 in the same or a similar trimeric complex to regulate the transcription of various targets involved in diverse biological processes.

NF-YA8 may acts as a hub integrating age and sugar signals

The influence of sugar on plant growth and development has become a hot topic in recent years (Moore et al., 2003; Smeekens et al., 2010). Based on the results of RT-qPCR, GUS staining of NF-YA8p::GUS, and bioluminescence activity analyses of NF-YA8p::YA8-LUC, we found that the transcript level and protein abundances of NF-YA8 were high in juvenile-stage leaves but dramatically decreased with increasing leaf age, consistent with the expression patterns of MIR156A (Fig. 6A–C). We further found that transcription of NF-YA8 was significantly repressed by sugar treatment (Fig. 6D, E), consistent with the expression patterns of MIR156A (Yang et al., 2013; Yu et al., 2013), indicating that sugar acts upstream of NF-YA8. In plants, endogenous sugar contents undergo diurnal changes, typically being high during daylight and low at night (Bläsing et al., 2005). Whether the transcript level or protein abundance of NF-YA8 also exhibits a rhythmic pattern remains an open question.

NF-Y is highly conserved from prokaryotes to eukaryotes (Edwards et al., 1998; Petroni et al., 2012) and plays important roles in diverse developmental processes (Hou et al., 2014; Siriwardana et al., 2014; Gnesutta et al., 2017; Hu et al., 2018; Luo et al., 2018; Hwang et al., 2019). Our transcriptome data in NF-YA8-3MYC plants suggested that many genes related to multiple aspects of growth, such as cell division and elongation, embryonic development, hormone responses, and stress responses, were significantly altered by NF-YA8, indicating its essential roles in regulating the transcription of various cellular processes (see Supplementary Fig. S6). Nevertheless, most of the differentially expressed genes and phenotypes in NF-YA8-3MYC transgenic plants have an amplified function caused by constitutive expression of NF-YA8 and might not resulted from the endogenous function of NF-YA8 alone. In this study, we connected the NF-YA8 subunit with the well-characterized miR156, which regulates the juvenile-to-adult transition and flowering time in plants (Fig. 7). In the juvenile phase leaves, NF-YA8 integrates developmental age and sugar signals, directly triggering miR156–SPLs-regulated processes and repressing the juvenile-to-adult transition. In the adult phase leaves, increased developmental age and accumulation of sugar represses the expression of NF-YA8 and releases the inhibition of flowering. NF-YA8 may also regulate the expression of other targets related to flowering time regulation to ensure flowering at the right development stage and season.

Fig. 7.

A model for NF-YA8 regulation of the juvenile-to-adult transition in Arabidopsis. Arabidopsis NF-YA8 directly binds to promoters of MIR156s and activates their expression in juvenile-stage leaves, thus repressing the vegetative phase transition. The transcription of NF-YA8 declines gradually as plant age increases and is rapidly repressed by sugar, thus releasing the repression of the vegetative phase transition.

Supplementary data

Supplementary data are available at JXB online.

Fig. S1. Identification of nf-ya8 mutants and NF-YA8 overexpression transgenic lines.

Fig. S2. Morphological phenotype analysis of NF-YA8-3MYC overexpression transgenic plants.

Fig. S3. NF-YA8 directly binds to the promoter regions of MIR156s.

Fig. S4. The transcription of MIR156s but not SPLs is directly activated by NF-YA8.

Fig. S5. The distributions of NF-Y binding sites in the promoter of MIR156A in multiple species.

Fig. S6. Transcriptome analysis of NF-YA8-regulated genes in Arabidopsis.

Table S1. A list of the primers used in this study.

Table S2. The expression levels of all the detected genes in NF-YA8-3MYC and WT No-0 plants.

Table S3. The expression levels of differentially expressed genes in NF-YA8-3MYC and WT No-0 plants.

Table S4. The expression levels of flowering time-related genes in NF-YA8-3MYC and WT No-0 plants.

Table S5. The expression levels of various flowering regulatory pathway-related genes showed in Fig. 2A in NF-YA8-3MYC and WT No-0 plants.

Acknowledgements

We thank Dr Haiyang Wang for kindly providing MIM156 seeds, and Dr Xingliang Hou for kindly providing the mutant seeds of NF-YA8 (CS876601 and CS873143). This work was supported by grants from the National Natural Science Foundation of China (31870266 and 31670249), Funds of the Shandong ‘Double Tops’ Program (to GL), and the China State Key Laboratory of Crop Biology (DXKT201706).

Author contributions

GL and HZ conceived the study; HZ and KL conducted the experiments; GL, HZ and KL drafted the manuscript. HZ, KL, LM, SG and QC edited the draft, and all authors approved the final version of the manuscript.

Conflict of interest

The authors declare no conflicts of interest.