PRMT6 physically associates with nuclear factor Y to regulate photoperiodic flowering in Arabidopsis

The timing of floral transition is critical for reproductive success in flowering plants. In long-day (LD) plant Arabidopsis, the floral regulator gene FLOWERING LOCUS T (FT) is a major component of the mobile florigen. FT expression is rhythmically activated by CONSTANS (CO), and specifically accumulated at dusk of LDs. However, the underlying mechanism of adequate regulation of FT transcription in response to day-length cues to warrant flowering time still remains to be investigated. Here, we identify a homolog of human protein arginine methyltransferases 6 (HsPRMT6) in Arabidopsis, and confirm AtPRMT6 physically interacts with three positive regulators of flowering Nuclear Factors YC3 (NF-YC3), NF-YC9, and NF-YB3. Further investigations find that AtPRMT6 and its encoding protein accumulate at dusk of LDs. PRMT6-mediated H3R2me2a modification enhances the promotion of NF-YCs on FT transcription in response to inductive LD signals. Moreover, AtPRMT6 and its homologues proteins AtPRMT4a and AtPRMT4b coordinately inhibit the expression of FLOWERING LOCUS C, a suppressor of FT. Taken together, our study reveals the role of arginine methylation in photoperiodic pathway and how the PRMT6-mediating H3R2me2a system interacts with NF-CO module to dynamically control FT expression and facilitate flowering time. Supplementary Information The online version contains supplementary material available at 10.1007/s42994-021-00065-y.


INTRODUCTION
Floral transition is controlled by several exogenous and endogenous developmental signals to ensure appropriate flowering time. To date, this transition is regulated by the photoperiodic, vernalization, autonomous, gibberellin, and the miR156-SPL module-mediated aging pathways (Wang et al. 2009;Andrés and Coupland 2012;He 2012;Bouché et al. 2017), in which the day length (photoperiod) is a stable seasonal signal to guarantee optimal flowering time in plants (Andrés and Coupland 2012;Romera-Branchat et al. 2014). Arabidopsis thaliana, a facultative long-day (LD) plant, can respond to photoperiodic signals to control floral transition by modulating the transcription of FLOWERING LOCUS T (FT) specifically in leaf vasculature via CON-STANS (CO)-dependent pathway (Andrés and Coupland 2012;Song et al. 2015). The B and C subunits of Nuclear Factor Y (NF-YB and NF-YC), as histone fold domain (HFD) NF-YB/NF-YC dimer, directly associate with CO thus compose of a trimeric NF-CO complex to regulate the FT expression (Wenkel et al. 2006;Gnesutta et al. 2018). Two NF-YB subunits (NF-YB2 and NF-YB3), and three NF-YC subunits  have been shown to interact with CO to activate FT expression (Kumimoto et al. 2010;Cao et al. 2014). Subsequently, FT protein, a major component of mobile florigen, transports from leaf veins to the shoot apex and leads to the floral transition (Andrés and Coupland 2012;Liu et al. 2013;Putterill and Varkonyi-Gasic 2016).
Except for the NF-CO complex, another regulator of FT expression is Polycomb group (PcG) which function to silence the transcription of FT (Wang et al. 2014;Mü ller-Xing et al. 2014;Luo et al. 2018). PcG proteins contain Polycomb repressive complex 1 (PRC1) and PRC2. PRC2 acts as a methyltransferase complex to catalyze histone 3 lysine-27 trimethylation (H3K27me3) on target chromatin, while PRC1 has been shown to maintain the H3K27me3 mark, leading to additional transcription repression by repressive chromatin modifications (Mozgova and Hennig 2015;Fö rderer et al. 2016). In Arabidopsis, two plant-unique BAH (Bromo adjacent homology) proteins SHORT LIFE (SHL) and EARLY BOLTING IN SHORT DAYS (EBS) form with a complex with EMBRYONIC FLOWER 1 (EMF1) to inhibit FT expression by reading the repressive H3K27me3 marks on FT locus (Li et al. 2018;Yang et al. 2018;Jing et al. 2019a). As another reader of H3K27me3, the LIKE HETEROCHROMATIN PROTEIN 1 (LHP1) has been shown to maintain this mark on FT locus (Turck et al. 2007;Zhang et al. 2007), and LHP1 can also directly interact with EMF1 to silence FT expression and thus to repress photoperiodic flowering (Wang et al. 2014). Remarkably, the NF-CO module is partly co-activated to relieve Polycomb repression on the transcription of FT, resulting in the de-repression of FT that confer the LD induction of floral transition in Arabidopsis (Luo et al. 2018). In addition, other chromatin modifications such as histone deacetylation SIN3 LIKEs (SNLs) and chromatin remodeler PICKLE (PKL) are involved in the regulation of flowering by facilitating FT expression at dusk (Huang et al. 2019;Jing et al. 2019b). Although these studies have emphasized the significance of epigenetic regulation in photoperiodic flowering, it is still to be established whether and how other chromatin modifiers 'communicate' with NF-CO and NF-Y in response to inductive LDs to modulate FT expression.
In this study, we identify a positive regulator of flowering PRMT6, which has been shown to catalyze the asymmetric dimethylation of R2 on H3 (H3R2me2a) (Guccione et al. 2007;Hyllus et al. 2007;Iberg et al. 2008). Further investigation finds that PRMT6 gene and its encoding protein accumulate in leaf veins at dusk, suggesting that PRMT6 may be associated with FT gene. PRMT6 interacts with NF-YCs and enhances their promotion on FT transcription. Moreover, AtPRMT6 and its homologues proteins AtPRMT4a and AtPRMT4b coordinately inhibit the expression of FLC, a suppressor of FT. Our results uncover the function of PRMT6 in plants, and provide insight into the role of arginine methylation in regulating photoperiodic flowering by 'communicating' with transcription factors.

PRMT6 interacts with NF-Y subunits
Two subunits of NF-Y proteins (NF-YBs and NF-YCs) have been reported to interact with the CO and form NF-CO complex to regulate photoperiodic flowering (Hou et al. 2014;Gnesutta et al. 2018). We observed that two NF-YC members, NF-YC3 and its homolog NF-YC9, could interact with PROTEIN ARGININE METHYLTRANSFER-ASE 6 (AT3G20020; PRMT6) in yeast cells (Fig. 1A, B). Additionally, PRMT6 also strongly interacted with NF-YB3 in yeast cells, but did not interact with NF-YB2 and NF-YC4 (Fig. S1a). Next, we performed in vivo bimolecular fluorescence complementation (BiFC) assay to verify these interactions. The enhanced yellow fluorescent protein (EYFP) with non-fluorescent N-terminal was fused to the full-length CDS of NF-YC3, NF-YC9, and NF-YB3, and the C-terminal fragments were fused to PRMT6. When PRMT6-cEYFP and NF-YC3-nEYFP were co-expressed in Arabidopsis mesophyll protoplasts, the fluorescence was observed in nucleus, but not from protoplasts co-expressing NF-YC3-nEYFP and cEYFP alone, or PRMT6-cEYFP and nEYFP alone (Fig. 1C). Similarly, we also confirmed the interactions of PRMT6 with NF-YC9 (Fig. 1D), and PRMT6 with NF-YB3 ( Fig. S1a, b). Concomitantly, a transient coimmunoprecipitation (Co-IP) assay was performed to testify the direct interaction of PRMT6 and NF-YC3 (Fig. S1c). We subsequently explored whether PRMT6 could physically associate with CO or FT protein, and found that PRMT6 did not interact with CO or FT in yeast cells (Fig. S1a).
PRMT6 enhances the promotion of NF-YC3;4;9 on flowering by facilitating FT expression Analyses of the GUS reporter expression in transgenic plants expressing GUS under control of the PRMT6 promoter indicated that PRMT6 promoter was active in the vascular bundle cells ( Fig. 2A). Subsequently, we measured the PRMT6 expression pattern every 4 h over a 24 h LD cycle in Col. PRMT6 mRNA abundance increased during daylight, peaks at ZT16 and then decreased (Fig. 2B). Additionally, we constructed a PRMT6-FLAG expression line driven by the native promoter region of PRMT6 and measured the protein abundance in the PRMT6-FLAG lines. PRMT6 protein varied diurnally and also peaked at ZT16 under LD conditions (Fig. 2C, D). Subcellular localization analysis showed that PRMT6 is a nuclear protein (Fig. S2). Collectively, nucleic protein PRMT6 displayed diurnal expression pattern and accumulated at dusk under LDs, which was consistent with the previously reported expression pattern of FT-GUS (Gu et al. 2013 indicated that PRMT6 might modulate flowering by regulating FT transcription. Next, two transfer DNA (T-DNA) insertion singlemutant prmt6-1 (Sail_385_A06) and prmt6-2 (Salk_151679C) (Figs. 3A; S3a) were used to explore its biological function in flowering. The total number of leaves of prmt6-1 and prmt6-2 mutant are not significantly different with Col under LD and SD conditions (Figs. 3B, C, S3b). To further confirm whether PRMT6 overexpression contributed to flowering time, we constructed overexpression lines and also found no significant differences between the Col and PRMT6 overexpression lines under LD conditions (Fig. S3c). We further explore whether PRMT6 mutation could affect the flowering phenotype of nf-yc3;4;9 mutants, and crossed prmt6-1 mutant with the nf-yc3;4;9 triple mutant to produce the quadruple mutant prmt6-1;nf-yc3;4;9. The total number of leaves in the prmt6-1;nf-yc3;4;9 was significantly more than that of nf-yc3;4;9 (Fig. 3B, D), suggesting that PRMT6 mutation could delay the flowering time of nf-yc3;4;9 mutant. Moreover, the expression of FT in prmt6-1;nf-yc3;4;9 mutant was lower than that of the nf-yc3;4;9 triple mutant line at ZT16 under LDs (Fig. 3E), in accordance with the additive action of prmt6-1 to delay flowering of nf-yc3;4;9.
PRMT6 directly binds on the FT locus and affects its H3R2me2a levels Because PRMT6 homologs have been identified to catalyze H3R2me2a in mammals (Guccione et al. 2007; methyltransferase (MTase) I, post-I, II, and III as well as the THW loop, which were involved in R methyltransferase activity (Fig. S4a, b). We then examined whether the PRMT6 mutation could affect the global H3R2me2a levels. Histones were extracted from 10-day-old seedlings and probed with antibodies against H3R2me2a and H3. The global levels of H3R2me2a in prmt6-1 were similar to these in Col (Fig. S5a). Next, we performed chromatin immunoprecipitation (ChIP) assays to explore whether PRMT6 could bind on FT locus and affect its H3R2me2a methylation level. First, we constructed a PRMT6 antibody and confirmed its specificity (Fig. S5b). ChIP assays were conducted using 10-day-old prmt6-1 and Col seedlings grown under LDs. PRMT6 was enriched at the FT promoter region at ZT16 but not at ZT8 in Col compared to prmt6-1 mutant (Fig. 4A). Considering the NF-YCs' transcript exhibits photoperiodic expression pattern under LD conditions (Fig. S6), we further performed ChIP experiments in Col, prmt6-1, nf-yc3;4;9, and prmt6-1;nf-yc3;4;9 lines at ZT8 and ZT16 of LDs, to verify whether PRMT6 could affect the H3R2me2a level to FT chromatin in the nf-yc3;4;9 mutant background. The H3R2me2a methylation level at FT locus of the prmt6-1; nf-yc3;4;9 quadruple mutant was significantly reduced at ZT16 but not at ZT8 (Fig. 4B). Collectively, these results demonstrated that PRMT6 mediated H3R2me2a modification to modulate FT transcription at the end of daylight. involved in vernalization and autonomous pathways (Niu et al. 2008). To clarify the genetic interaction between PRMT6 and PRMT4a/4b in regulating floral transition, we then generated prmt6-1;4a;4b and prmt6-2;4a;4b triple mutants by crossing. And the homozygous prmt6-1;4a;4b were used for subsequent analysis. The total leaf number of the prmt6-1;4a;4b and prmt6-2;4a;4b was more than prmt4a;4b double mutant, indicating that prmt6;4a;4b triple mutant exhibited the delayed flowering phenotype compared to prmt4a;4b double mutant under LD conditions (Fig. 5A-C). Next, we generated prmt6-1;4a;4b;ft-10 quadruple mutant lines to determine whether the ft-10 mutation could rescue the prmt6-1;4a;4b triple mutant phenotype. As expected, we found that the quadruple mutant line displayed a similar number of total leaves to the ft-10 single mutant (Fig. 5D), suggesting that the genetic mechanism regulating the prmt6-1;4a;4b phenotype could be completely rescued by the ft-10 mutation and that FT could thus be considered their downstream gene. Further investigation showed that the expression level of FLC in prmt6;4a;4b triple mutant was A Phenotypes of Col, prmt6-1, prmt4a;4b, and prmt6-1;4a;4b mutants. Plants were grown under LD conditions. The white arrow indicates that the prmt4a;4b double mutants were bolting. B Flowering time of Col, prmt6-1, prmt4a;4b, and prmt6-1;4a;4b under LD conditions. More than 15 plants for each line were scored; bars indicate the s.d.; significant differences between the means of the indicated genotypes were revealed by a two-tailed Student's t test (n.s., no significant difference; *p \ 0.05; **p \ 0.01). C Flowering time of Col, prmt6-2, prmt4a;4b, and prmt6-2;4a;4b grown under LD conditions. D Flowering time of Col, prmt6-1;4a;4b, ft-10, and prmt6-1;4a;4b ft-10 grown under LD conditions. E, F the relative transcription levels of FLC (E) and FT (F) in seedlings of the indicated genotypes as quantified by RT-qPCR. The transcription levels were normalized to UBQ10, and the fold changes relative to Col are presented Ó The Author(s) 2021, corrected publication 2021 aBIOTECH (2021) 2:403-414 significantly higher than that in prmt4a;4b double mutant at ZT8 and ZT16 under LD conditions (Fig. 5E). PRMT6 mutation only inhibited the FT mRNA abundance in prmt6;4a;4b triple mutant at ZT16 under LD conditions (Fig. 5F). Together, these findings revealed that PRMT6 may be not only involved in regulation of photoperiodic FT expression through NF-CO module, but also exhibits redundancy with PRMT4a/PRMT4b to regulate FLC expression, thus to promote floral transition in Arabidopsis.

DISCUSSION
PRMT6 serves as co-factor of NF-YCs and PRMT4s during flowering In animal systems, PRMT6 has been reported to play an important role in the regulation of disease and developmental processes (Boulanger et al. 2005;Yang and Bedford 2013), and acts as writers to catalyze H3R2me2a (Guccione et al. 2007;Hyllus et al. 2007;Iberg et al. 2008). However, the function of PRMT6 of regulating plant development remains unclear. In this study, we have first illustrated the involvement of AtPRMT6, an Arabidopsis homolog of human HsPRMT6, in the flowering transition. In Arabidopsis prmt6 mutants exhibited a similar phenotype to wild type, whereas PRMT6 mutation further delays the flowering time of prmt4a;4b (Fig. 5;Niu et al. 2008), suggesting PRMT6 could function redundantly with PRMT4a/ PRMT4b in controlling floral transition. In addition, our results indicated that the PRMT6 mutation also strengthened the late flowering phenotype of nf-yc3;4;9 (Fig. 3). These results demonstrated PRMT6, which might serve as a co-factor of NF-YCs and PRMT4s, synergistically modulated floral transition.
PRMT6 control floral transition via NF-Y-CO module or FLC-dependent pathway In Arabidopsis, movement of the FT protein (also known as florigen) contributes to inducing the flowering transition through long-distance signaling from the leaf vascular tissue (phloem) to the SAM (Corbesier et al. 2007), and FT-GUS was mainly expressed in the leaf vascular tissues (Gu et al. 2013). Our study found that PRMT6 exhibited a similar expression pattern to that of FT in the leaf phloem ( Fig. 2A, B). Moreover, the expression patterns of Flag-PRMT6 also showed rhythmic oscillations under LD conditions and accumulated at dusk under LDs (Fig. 2C), coinciding with the peak in CO protein level and increasing FT expression. These results demonstrated that PRMT6 could regulate CO-FT module. The nuclear factor NF-Y could bind to target DNA sequences accumulated on silent chromatin regions and act as a 'pioneer' to open up the chromatin structure to activate gene expression (Fleming et al. 2013;Oldfield et al. 2014). In addition, its subunits have been reported to interact with the CO protein (forming NF-CO complexes) to regulate flowering time (Gnesutta et al. 2018). In our study, PRMT6 was associated with NF-YC proteins, but did not interact with CO in yeast cells (Fig. S1). Considering the interaction between NF-YC proteins with CO testified by the previous studies (Hou et al. 2014;Gnesutta et al. 2018), we suppose that PRMT6 could interplay with CO by NF-YC proteins. Moreover, we confirmed the occupancy of PRMT6 at FT loci, as well as the reduced H3R2me2a modification in prmt6-1;nf-yc3;4;9 compared with nf-yc3;4;9 (Fig. 4). These data demonstrated that NF-YC proteins could recruit PRMT6 to FT promoter, consequently catalyzing H3R2me2a modification and finally accelerate floral transition. On the other hand, we observed PRMT6 and PRMT4 proteins synergistically inhibited the expression of FLC (Fig. 5). Taken together, PRMT6, NF-YCs, and PRMT4s synergistically modulated floral transition by CO-FT module or FLC-related pathway.
In conclusion, our study revealed that AtPRMT6, a PRMT6 homolog in Arabidopsis, acts as a positive regulator of floral transition. AtPRMT6 physically associates with three NF-Y subunits to bind to the FT locus around dusk (ZT16) of LDs, consequently changing the Fig. 6 A working model for PRMT6 promoting flowering time in Arabidopsis. AtPRMT6 physically associates with three NF-Y subunits to bind to the FT locus around dusk (ZT16) of LDs, consequently changing the methylation abundance on FT locus to promote its expression at ZT16 of LDs. In addition, PRMT6 play a redundant role with PRMT4a/PRMT4b via inhibiting FLC expression during floral transition Ó The Author(s) 2021, corrected publication 2021 methylation abundance on FT locus to promote its expression at ZT16 of LDs. In addition, PRMT6 play a redundant role with PRMT4a/PRMT4b via inhibiting FLC expression during floral transition (Fig. 6). Our study reveals the role of arginine methylation in photoperiodic pathway and how the PRMT6-mediating H3R2me2a system interacts with NF-CO module to dynamically control FT expression and facilitate flowering time.

MATERIALS AND METHODS
Plant materials, growth conditions, and flowering time

RNA extraction and RT-qPCR
Total RNA was extracted from the aerial parts of 10-dayold seedlings grown under LD conditions using the Qiagen RNeasy Plus Mini Kit and then reverse-transcribed into cDNA with M-MLV reverse transcriptase and oligo (dT) primers (Promega). Real-time quantitative PCR (RT-qPCR) was performed using the Roche Light Cycler 480 II System using SYBR Green PCR master mix, as described previously (Gu et al. 2013). Each sample was quantified at least three times and normalized to UBQ10. The relative transcription levels was calculated as 2 -4Ct (Gu et al. 2013). The primer pairs for FT, NF-YC3, NF-YC9, TUB2, and UBQ10 amplification were described previously (Gu et al. 2013;Hou et al. 2014).

Plasmid construction and plant transformation
To generate the pPRMT6-PRMT6:FLAG plasmid, the fulllength PRMT6 CDS without the stop codon was first fused with a 3 9 FLAG tag and cloned into pHGW vector via Gateway technology. The sequences of the primers used for plasmid construction are specified in Table S1. Agrobacterium tumefaciens GV3101 was used to generate stable transgenic lines.

Histochemical b-glucuronidase staining
To construct AtPRMT6-GUS, 4.5 kb genomic fragments including the 2.4 kb promoter and 2.1 kb genomic coding sequence of PRMT6 (including exons and introns) were inserted into the pMDC162 vector via Gateway technology (Invitrogen). Histochemical b-glucuronidase (GUS) staining was performed as described previously (Gu et al. 2013). Briefly, GUS staining was carried out with 5-and 10-day-old seedlings after hygromycin selection by immersing them in X-Gluc (5bromo-4-chloro-3-indoyl b-D-glucuronide) (0.5 mg/ml) (Gold Biotechnology; USA), followed by vacuum briefly to ensure uniform staining. The tissues were then incubated in X-Gluc at 37°C for various times followed by incubation in 70% ethanol at 37°C overnight to remove the chlorophyll from the green tissues.

Subcellular localization
For PRMT6-GFP fusion protein construction, the 1.3 kb CDS (coding sequence) region containing all the exons of PRMT6 was inserted between the 35S promoter and GFP in the pMDC85-GFP vector via Gateway technology (Invitrogen). The subcellular localization of PRMT6 was conducted as described previously (Gu et al. 2013). GFP fluorescence signals were observed and recorded using a Zeiss LSM 700 confocal laser scanning microscope.

Preparation of polyclonal antibody
The synthetic PRMT antibody was performed as described previously (Geng et al. 2020;Zhang et al. 2021). Briefly, a 15-amino acid fragment of PRMT6 (residues 105-119; TYREAIMQHQSLIEG) was synthesized by the solid-phase peptide synthesis (SPPS) method and confirmed by mass spectrometry (MS) and high-performance liquid chromatography (HPLC). The polypeptide was coupled through the Sulfo-SMCC agent to react with keyhole limpet hemocyanin (KLH) for preparation of the immune complex. The prepared immune complex was immunized to two rabbits to produce the antibody and then purified by a specific affinity chromatography column. The synthesized antibody was verified by western blot analysis using total proteins extracted from wild-type and prmt6-1 plants.

Histone extraction and immunoblotting
Histone protein extraction and western analysis were performed as described previously (Zhang et al. 2021). Briefly, total histones were extracted from 10-day-old seedlings grown in LD conditions, separated on an SDS-PAGE gel, and subsequently transferred to a 0.2-lm nitrocellulose membrane (Bio-Rad). The protein blots were probed with anti-H3 (Abcam, Cat#: ab1791) and H3R2me2a (Abcam, Cat#: ab175007). Immunoblotting was visualized by chemiluminescence. Blotting signals were captured using ImageJ software, and the relative protein level of H3R2me2a was normalized to that of H3. Experiments were repeated at least two biological times.
Yeast two-hybrid (Y2H) assay Yeast two-hybrid assays were conducted using the Matchmaker GAL4 Two-Hybrid System 3 (Clontech) as per the manufacturer's instructions. The full-length coding sequences of PRMT6, NF-YC3, NF-YC4, NF-YC9, NF-YB3, CO, and FT were cloned into the pGADT7 and pGBKT7 vectors and subsequently introduced into the yeast strain AH109. Yeast cells were spotted on selective media lacking leucine (L), tryptophan (W), histidine (H), and adenine (A) for interaction detection or on drop-out media lacking L and W (as control).

Co-immunoprecipitation (Co-IP) assay
Co-immunoprecipitation (Co-IP) experiments were carried out as previously described with some modifications (2013). Briefly, plasmid pairs were performed by a transient expression system using Arabidopsis mesophyll protoplast (Yoo et al. 2007). After 12-18 h incubation, total proteins were extracted from Arabidopsis mesophyll protoplasts and immunoprecipitated with anti-FLAG M2 affinity gel (Sigma, Cat#: A2220), and the immunoprecipitated protein was detected by western blotting with anti-FLAG (Sigma, Cat#: A8592).

Chromatin immunoprecipitation (ChIP) and ChIP-qPCR analysis
ChIP experiments were carried out as previously described (Gu et al. 2013). Briefly, total chromatin was extracted from 10-day-old seedlings grown under LD conditions and immunoprecipitated with anti-H3R2me2a (Abcam, Cat#: ab175007) and anti-PRMT6 (synthesized in this study). Quantitative PCR (qPCR) was conducted to measure the amounts of FT and the constitutively expressed TUB8 fragments on a Roche Light Cycler 480 II System using SYBR Green PCR master mix. The ChIP-qPCR primer pairs for FT, were described previously (Gu et al. 2013).
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/.