Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jul 1.
Published in final edited form as: New Phytol. 2021 May 2;231(1):182–192. doi: 10.1111/nph.17366

DEK Domain-Containing Proteins Control Flowering Time in Arabidopsis

Wei Zong 1, Bo Zhao 1, Yanpeng Xi 1, Yogendra Bordiya 1, Hyungwon Mun 1, Nicholas A Cerda 1, Dong-Hwan Kim 1, Sibum Sung 1
PMCID: PMC8985477  NIHMSID: NIHMS1792938  PMID: 33774831

Introduction

Floral transition is a critical developmental change during the life cycle of plants. Thus, flowering pathways are sophisticatedly controlled by both internal and external cues. In Arabidopsis, four major flowering pathways have been defined: photoperiod, autonomous, gibberellin (GA), and vernalization, all of which eventually control two major floral integrator genes, FLOWERING LOCUS T (FT) and SUPPRESSOR OF OVEREXPRESSION OF CONSTANTS 1 (SOC1) (Kim et al., 2009; Kim & Sung, 2014; Whittaker & Dean, 2017). The levels of expression of SOC1 and FT are partially regulated by the major flowering repressor FLOWERING LOCUS C (FLC) and its five homologs, MADS AFFECTING FLOWERING 1 (MAF1) to MAF5 (Luo & He, 2020). Chromatin regulators that control the expression of FLC and MAFs have been well documented (Kim & Sung, 2014; Whittaker & Dean, 2017). For example, Polycomb-repressive complex 2 (PRC2) deposits a repressive histone mark via the trimethylation of Histone H3 Lysine 27 (H3K27me3) at FLC to repress its expression (He, 2012; Luo & He, 2020). On the other hand, trimethylation of Histone H3 lysine 36 (H3K36me3) is catalyzed by EARLY FLOWERING IN SHORT DAYS (EFS) and correlates with active FLC transcription (He, 2012; Luo & He, 2020). Studies on the chromatin-level regulation of FLC and its homologs have served as an excellent model of epigenetic regulation in plant development (Whittaker & Dean, 2017).

DEK domain-containing protein (DEK) is an abundant DNA-binding protein that was originally identified as a DEK-NUP213 fusion protein in human acute myeloid leukemia (AML) (von Lindern et al., 1992; de Albuquerque Oliveira et al., 2018). Therefore, DEK is an oncoprotein and has been extensively studied in humans. In mammals, DEK has been implicated in DNA replication, DNA repair, mRNA splicing, and gene regulation (Alexiadis et al., 2000; Soares et al., 2006; Kavanaugh et al., 2011). However, DEK’s role in the regulation of gene expression still remain elusive, in part because DEK proteins are involved in both transcriptional activation and transcription repression (Sanden & Gullberg, 2015). For example, DEK localizes at euchromatin regions and associates with active histone modifications, such as trimethylation of Histone H3 Lysine 3 (H3K4me3) (Sawatsubashi et al., 2010; Ivanauskiene et al., 2014). DEK is also considered to be a “gatekeeper” of chromatin because DEK controls the distribution of Histone H3.3 by restricting the access of dedicated chaperones (Sawatsubashi et al., 2010; Ivanauskiene et al., 2014). Furthermore, DEK also interacts with transcriptional activators such as AP2α, C/EBPα and MLLT3 and promotes their transcriptional activation (Campillos et al., 2003; Shibata et al., 2010; Koleva et al., 2012). On the other hand, DEK is also necessary for heterochromatin integrity by directly interacting with Heterochromatin Protein 1α and enhancing its binding to H3K9me3 (Kappes et al., 2011). Therefore, DEK serves as a chromatin architectural modulator and affects chromatin structure through various routes.

In plants, the biological function of DEK proteins has been reported in a few studies (Waidmann et al., 2014; Anna Brestovitsky, 2019). The Arabidopsis genome contains four DEK genes: DEK1, DEK2, DEK3, and DEK4 (Waidmann et al., 2014). Among them, DEK3 is the only DEK that has been studied in detail in Arabidopsis (Waidmann et al., 2014; Anna Brestovitsky, 2019), in part because DEK3 is expressed at the highest level among the four DEK proteins (Waidmann et al., 2014). In Arabidopsis, DEK3 associated with the Topoisomerase Iα (Top1α) in vivo and DEK3 can change the topology of DNA in vitro (Waidmann et al., 2014). Moreover, DEK3 can affect nucleosome occupancy and chromatin accessibility to repress gene expression at some loci (Waidmann et al., 2014). DEK3 loss-of-function mutations altered stress tolerance to both salt and high-temperature in Arabidopsis (Waidmann et al., 2014). A recent study also showed that DEK2, DEK3, and DEK4 associate with H2A.Z in vivo and DEK3 may play a role in balancing the response between growth and arrest via H2A.Z-nucleosome in Arabidopsis (Anna Brestovitsky, 2019).

In this study, we investigate the biological roles of DEK3 and DEK4 and find that they redundantly control the floral transition in Arabidopsis. We show that DEK3 and DEK4 are necessary for the active transcription of related floral repressor genes, FLC, MAF4, and MAF5. DEK3 and DEK4 directly associate with FLC, MAF4, and MAF5 loci to facilitate their active transcription of these related floral repressor genes to prevent precocious flowering in Arabidopsis.

Materials and Methods

Plant materials and growth conditions

Arabidopsis thaliana (L.) accession Columbia (Col-0), dek1-2 (SALK_113239), dek2-3 (SALK_137152), dek3-2 (SALK_112581), and dek4-2 (SALK_060488) lines were obtained from Arabidopsis Biological Resource Center (Columbus, OH). Primers used for genotyping (Listed in Supplemental Table S1) were synthesized by Millipore Sigma. FRI-Col, flc-3, FLC:GUS, fld-3 lines were described previously (Lee et al., 1994; Michaels & Amasino, 1999; He et al., 2003; Michaels et al., 2005; Yoo et al., 2005). Plants were grown in controlled environmental chambers with cool white fluorescent lights and maintained at 22 °C. The photoperiodic cycle was 16h light/ 8h dark (long day, LD) or 8h light/ 16h dark (short day, SD). For vernalization treatments, seeds were surface sterilized, germinated on ½ MS medium, and grown for 5 days at 22°C; subsequently, seedlings were transferred to 4°C for indicated times under SD.

Plasmid construction and transformation

To generate pGWB16-DEK3::DEK3-cMYC vector, a fragment containing a 2.1-kb promoter region and the 4-kb genomic fragment of DEK3 was amplified with primers DEK3_pro_F/DEK3-pENTR_R (Supplemental Table S1) and cloned with pENTR™/D-TOPO™ Cloning Kit (Invitrogen) into pENTR and further cloned into the binary vector pGWB16 by using Gateway™ LR Clonase™ II Enzyme mix (Invitrogen). The pGWB16-DEK4::DEK4-cMYC vector was generated with the same method, and the primers used were listed in Supplemental Table S1. To generate 35S::DEK3-GFP and 35S::DEK4-GFP, DEK3 and DEK4 coding sequences were amplified with primers DEK3-pENTR_F/R and DEK4-pENTR_F/R (Supplemental Table S1), respectively, and cloned into pENTR, which further cloned into pEH18 through Gateway™ LR reaction. The constructs above were transformed into Agrobacterium Tumefaciens cells (GV3101) and then stably transformed into Arabidopsis using the floral dip method (Clough & Bent, 1998).

mRNA expression analysis

Total RNA was extracted from 7-day-old seedlings using PureLink™ Plant RNA Reagent (Invitrogen). For RT-qPCR, 1μg total RNAs were treated with DNase I (Invitrogen) before reverse transcription, and then the first-strand cDNA was synthesized by M-MLV (Invitrogen). Real-time qPCR was performed on the ViiA 7 Real-Time PCR System (ABI) with AzuraQuant™ Green Fast qPCR Mix (Azura genomics). Each sample was quantified in triplicate. Primers used were listed in Supplemental Table S1. The constitutively expressed PP2A (AT1G69960) was used as an internal control for normalization.

RNA-Seq analysis

Total RNA was extracted from 7-day-old seedlings using PureLink™ Plant RNA Reagent (Invitrogen) according to the manufacturer’s instructions. 500ng RNA was used for the construction of the RNA-Seq libraries. The libraries were constructed using NEBNext® Ultra™ Directional RNA Library Prep Kit for Illumina® (E7420) according to the manufacturer’s instructions. The libraries were validated using Agilent Bioanalyzer and sequenced on an Illumina NEXTSeq500 platform by the Genomic Sequencing and Analysis Facility at the University of Texas at Austin. After sequencing, reads were mapped to the TAIR 10 Arabidopsis genome using HISAT2 with default parameters (Kim et al., 2015). Differentially expressed gene were identified by using edgeR with default parameters. Genes with at least 1.5-fold change in expression and p Value < 0.05 between mutant and WT (Col-0) were considered to be differentially expressed.

Chromatin immunoprecipitation

ChIP experiments were performed as previously reported (Qian et al., 2018) with some modifications. Briefly, one gram of 7-day-old seedlings grown at LD was ground into fine powder in liquid nitrogen and cross-linked in nuclei isolation buffer I (10mM HEPES, pH 8.0, 1M Sucrose, 5mM KCl, 5mM MgCl2, 5mM EDTA, 0.6% Triton X-100, 0.4mM PMSF, and 1X protease inhibitor cocktail tablet (Thermo Scientific™) with 1% formaldehyde for 20 mins at room temperature. Crosslinking was stopped by adding glycine to a final concentration of 125mM. The homogenate was filtered through two-layer miracloth (Millipore), and then nuclei were pelleted and washed by ChIP buffer 2 (10mM Tris-HCl, pH 8.0, 0.25M Sucrose, 10mM MgCl2, 1% Triton X-100, 1mM EDTA, 5mM β- mercaptoethanol, and 1X protease inhibitor). Finally, the nuclei pellet was resuspended with Nuclear lysis buffer (50mM Tris-HCl, pH 8.0, 10mM EDTA, 1% SDS, 0.1mM PMSF, 1X protease inhibitor) and incubated on ice for 10 mins. Chromatin was sheared by Bioruptor® to an average size of about 500bp. Subsequently, sheared chromatin was further diluted by ChIP dilution buffer and incubated with 4μg cMYC (A3120, Santa Cruz) or H3K4me3 (ab8580, Abcam) or H3K27me3 (07–449, Millipore) or H3K36me3 (ab9050, Abcam) or RNA Pol II Ser5P (ab5131, Abcam) antibodies for overnight at 4°C. The samples were then incubated with 40μl Dynabeads™ Protein G (Thermo Fisher) for 2 h at 4°C. The beads were washed with low salt buffer, high salt buffer, LiCl buffer, and TE buffer. The DNA-protein complex was eluted by Elution buffer and reverse cross-linked at 65°C overnight. After proteinase K and RNase A treatments, DNA was purified by QIAquick® PCR Purification Kit and was used for quantitative PCR (Primers were listed in Supplemental Table S1).

Protein extraction and co-immunoprecipitation

Total proteins were extracted from 7-day-old seedlings using 2X Laemmli buffer (100mM Tris-HCl, pH 6.8, 4% SDS, 100mM DTT, 20% Glycerol, 2% β- mercaptoethanol). For Co-IP, 0.5 gram total proteins were extracted from 7-day-old seedlings using Plant Extraction buffer (25mM Tris-HCl, pH 7.5, 1mM EDTA, 150mM NaCl, 10% Glycerol, 1mM PMSF and 1X protease inhibitor). GFP-Trap® Dynabeads (Chromotek) was used to precipitate GFP tagged DEK3 and DEK4 protein. Subsequently, histone H3 was detected by western blotting using anti-H3 (ab1791, Abcam), GFP tagged DEK3 and DEK4 were detected with anti-GFP (ab290, Abcam), RNA Polymerase II Ser5P was detected with Anti-RNA polymerase II CTD repeat YSPTSPS (phospho S5) antibody (ab5131, Abcam).

Histochemical β-glucuronidase staining

8-old-seedlings were incubated with GUS staining buffer (1 mM EDTA, 0.2% Triton X-100, 2 mM potassium ferrocyanide, 2 mM potassium ferricyanide, 100 mM NaH2PO4, 100 mM Na2HPO4 pH 7.0, and 2mM 5-bromo- 4-chloro-3-indolyl-β-d-glucuronic acid (X-Gluc)) at 37°C for 8–10 h. After staining, seedlings were cleared by incubation in 70% ethanol for several hours. Stained seedlings were photographed using Stemi 508 Stereo microscope (Carl Zeiss).

MNase Assay

MNase assay was performed as previously reported () with some modifications. Briefly, the nuclei were isolated by following the ChIP assay as described above. The isolated nuclei were washed and digested with 8 unit/μl (final concentration) of Micrococcal Nuclease (NEB, M0247S) for 2 mins, 10 mins, 20 mins and 30 mins in digestion buffer at 37°C. The digestion was stopped by the stop buffer (1mM EGTA, 10mM EDTA, pH 8.0, 0.1% SDS) at 65°C for 30 mins. Reverse crosslinking was performed by overnight treatment at 65°C, DNA was recovered after Proteinase K and RNase A treatment by phenol/chloroform extraction and ethanol precipitation. qPCR was used subsequently to measure the DNA recovered from MNase digestion as compared with that in the control without MNase digestion.

Results

DEK3 and DEK4 redundantly inhibit flowering

There are four DEK genes in Arabidopsis. To address the function of DEK genes in plant development, we isolated loss of function mutants of DEK1~4 (Fig. 1a, Fig. S1a). There is no detectable transcript of corresponding DEK genes in these mutants (Fig. S1b–e). Although there is no observable vegetative growth defect, both dek3-2 and dek4-2 mutants exhibited significantly early flowering compared to the wild-type plants under both long-day (LD) and short-day (SD) conditions (Fig. 1b–d). However, we did not observe any difference in flowering time in dek1-2 and dek2-3 mutants (Fig. S1f), indicating that the biological function of DEK genes may differ among the members. To examine whether DEK3 and DEK4 have redundant roles in flowering, we crossed dek3-2 and dek4-2 to create dek3-2;dek4-2 (dd-2) double mutants. The dd-2 double mutant flowers much earlier than either dek3-2 or dek4-2 single mutants under both LD and SD (Fig. 1b–d). Therefore, DEK3 and DEK4 control floral transition in various growth conditions.

Fig. 1.

Fig. 1

Open in a new tab

DEK3 and DEK4 control flowering time in Arabidopsis. (a) Schematic diagram showing the gene structure of DEK3 and DEK4 and the location of T-DNA insertion sites in dek3-2 and dek4-2 mutants. Boxes represent exons and untranslated regions are indicated by gray boxes. Black lines indicate introns and other genomic sequences. (b) A representative image shows that dek3-2, dek4-2, and dd-2 mutants display early-flowering phenotype under LD. (c) and (d) The number of rosette leaves at flowering for dek3-2, dek4-2, and dd-2 mutants grown in LD (c) and SD (c). Error bars: ± s.d. (n = 12). (e) and (f) The early flowering phenotype of dek3-2 and dek4-2 were rescued by gDEK3-cMYC (e) and gDEK4-cMYC (f) under LD. Error bars: ± s.d. (n = 12). One-way ANOVA with Tukey’s multiple comparison test was conducted; asterisks indicate P < 0.01 of distinct groups.

To further confirm whether the early flowering phenotype of dek3-2 is due to the loss of DEK3, we transformed dek3-2 with the genomic fragment of DEK3 (Fig. 1e). We also confirmed the genetic redundancy of DEK3 and DEK4 in controlling flowering time by transforming dd-2 with the genomic fragment of DEK3 or DEK4. For each transformation, two independent homozygous lines were selected for further analysis. Both DEK3 and DEK4 transgenes could reverse the early flowering of dd-2 (Fig. 1f and Fig. S2a). The expression of DEK3 and DEK4 transgenes in complementation lines were further confirmed by protein blot analysis (Fig. S2b).

Transcriptome analysis of dek mutants

To molecularly characterize flowering behaviors of dek mutants, we investigated the gene expression profiles in dek3-2, dek4-2, and dd-2 plants by employing RNA-Seq analysis. A total of 287, 480, and 812 differentially expressed genes (DEGs) were identified in dek3-2, dek4-2, and dd-2 mutants, respectively (Fig. 2a). Gene ontology (GO) analysis showed that significant GO-term enrichments of DEGs in dd-2 mutants include pathways involved in cell wall organization, response to stimuli, and response to oxidative stress (Fig. 2b). However, we did not observe similar GO term enrichment in DEGs of either dek3-2 or dek4-2, indicating the functional redundancy between DEK3 and DEK4.

Fig. 2.

Fig. 2

Open in a new tab

Transcriptome analysis of dek mutants in Arabidopsis. (a) Venn diagram showing overlapped DEGs between dek3-2, dek4-2, and dd-2 mutants. (b) Heat map showing the gene ontology categories of DEGs from dek3-2, dek4-2, and dd-2 mutants. (c) Volcano plot displaying significantly up-regulated and down-regulated genes in dd-2 mutants compared to WT (Col-0), and the y axis is the -Log10 of the P-value for the significance of differential expression. (d) Quantification of the expression levels of FLC and MAFs in Col-0, dek3-2, dek4-2, and dd-2. Expression levels were shown as in counts per million reads (CPM). Error bars: ± s.d. (n = 2).

We then focused on DEGs in dd-2 for further analysis because dd-2 double mutants flower much earlier than either dek3-2 or dek4-2. Among DEGs in dd-2 double mutants, 272 genes were significantly down-regulated, whereas 540 genes were significantly up-regulated compared to wild-type (Fig. 2c). Interestingly, FLC is among 272 down-regulated genes in dd-2 (Fig. 2c). Furthermore, we found that two FLC homologs, MAF4, and MAF5, are also significantly down-regulated in dd-2 (Fig. 2c). However, three other FLC homologs, MAF1, MAF2, and MAF3, were not affected in these mutants (Fig. 2d). Consistent with the flowering time phenotypes among dek mutants, the abundance of transcripts of FLC, MAF4 and MAF5 is lower in each dek3-2 and dek4-2 single mutants, and they are even further down in dd-2 mutants in RNA-Seq analysis (Fig. 2d). Therefore, it is likely that the reduced expressions of FLC, MAF4, and MAF5 are responsible for the earlier flowering in dek mutants.

Analysis of flowering genes in dek mutants

RNA-Seq analysis indicates that the expression levels of FLC and its two homologs are reduced in dek mutants. We confirmed the results from RNA-seq analysis by reverse-transcription followed by quantitative PCR (RT-qPCR) (Fig. S3a). Consistent with the early flowering phenotype and RNA-seq analysis, FLC expression was decreased in both dek3-2 and dek4-2 single mutants and was even lower in dd-2 mutants (Fig. S3a). Moreover, the level of unspliced FLC transcripts also is lower in dd-2 mutant than Col-0 (Fig. S3b). In addition, complementation lines with either DEK3 or DEK4 could successfully recover the expression of FLC to a comparable level to that in the wild type (Fig. 1e and 1f, Fig. S3c and S3d), confirming the genetic redundancy between DEK3 and DEK4 in the activation of FLC. We chose two independent complementation lines for gDEK3-cMYC, gDEK3#6 and gDEK#11. Although both complementation lines could complement dek3-2 mutants, the transgene expression levels are different, as shown by protein blots (Fig. 1e and Fig. S2b). Indeed, the higher expression of transgene resulted in later flowering, reflecting the more robust activation of FLC (Fig. 1e). Similarly, the down regulation of MAF4 and MAF5 in dd-2 mutants was also successfully recovered by the introduction of wild-type copy DEK3 or DEK4 transgenes (Fig. S3e–h). Consistent with early-flowering phenotype under SD, the levels of FLC, MAF4 and MAF5 decreased in dek3-2 and dek4-2 single or double mutants but not in dek1-2 and dek2-3 mutants under SD (Fig. S4a–c). An apparent down-regulation of FLC in dd-2 mutants was also observed using FLC: GUS transgenic lines created in dd-2 mutants compared to those in the wild-type (Michaels et al., 2005)(Fig. S5a). In contrast, mRNA expression of one of the floral integrators, FT, was significantly increased in dd-2 mutants but then could be recovered in complementation lines (Fig. S5b and S5c).

Vernalization triggers the repression of FLC and other homologs in a quantitative manner in Arabidopsis (Kim & Sung, 2014). Because the early flowering observed in dek mutants is likely due to the reduced levels of FLC and its homologs, we further explored whether dek mutants compromise vernalization response. To address this, we first crossed dd-2 mutants into a vernalization-responsive Col-FRIGIDA (FRI) background, in which a functional FRI has been introgressed into Col-0 (Lee et al., 1994). The dd-2 in Col-FRI (ddF) flowered more rapidly than the Col-FRI wild-type plant, even without vernalization (Fig. S6a). However, the flowering of ddF plants was further accelerated by vernalization to comparable degrees observed in Col-FRI wild-type plants (Fig. S6a). In addition, the level of FLC is consistently lower in ddF mutants compared to the wild-type Col-FRI throughout vernalization (Fig. S6b). The vernalization-induced expression of VIN3 (Sung & Amasino, 2004) was not compromised in ddF mutants (Fig. S6c), confirming that DEKs are not involved in the vernalization response.

The dd-2 mutant partially suppresses the late-flowering of Col-FRI (Fig. S6a), suggesting that DEKs are required for the full FLC activation mediated by the FRI complex (Li et al., 2018). To test whether the effects of mutations in DEK genes are specific to the FRI-mediated FLC activation, we introgressed dd-2 mutations into an autonomous pathway mutant, fld-3 (He et al., 2003). Like ddF mutants, dd-2;fld-3 triple mutants flower earlier than fld single mutants (Fig. 3a), although dd-2 fld-3 triple mutants did not completely reverse the flowering time levels back to those of the wild-type Col-0. Partial suppression of late-flowering phenotypes and corresponding levels of FLC expression in both Col-FRI and fld-3 mutant by dd-2 mutations (Fig. 3b) suggest that DEK proteins function to ensure the full extent of transcriptional activation of FLC, independent of either FRI complex or autonomous pathway proteins.

Fig. 3.

Fig. 3

Open in a new tab

Decreased expression of FLC, MAF4 and MAF5 caused early flowering in dd-2 in Arabidopsis. (a) Flowering time of Col-0, dd-2, fld-3, dd-2;fld-3 under LD. (b) RT-qPCR analysis of FLC expression in indicated genotypes. (c) Flowering time of Col-0, dd-2, flc-3, dd-2;flc-3 under LD. Flowering time was measured by the number of rosette leaves at bolting. Error bars: ± s.d. (n=12). One-way ANOVA with Tukey’s multiple comparison test was conducted; asterisks indicate P < 0.05 of distinct groups. (d) to (f) RT-qPCR analysis of MAF4 (d), MAF5 (e), and FT (f) expression in indicated genotypes. Transcript levels were normalized to PP2A. Asterisks indicate statistically significant differences in the indicated plant pairs (two-tailed paired Student’s t-test, P < 0.001). Error bars: ± s.d. (n =2).

DEK3 and DEK4 are necessary for the transcriptional activation of floral repressors, FLC, MAF4, and MAF5

To genetically determine whether the decreased level of FLC expression is required for the early flowering, we introduced the dd-2 mutation into the flc-3 mutant by genetic crossing. As expected, the flowering time of the dd-2;flc-3 triple mutant flower similarly to dd-2 mutants but earlier than flc-3 mutants (Fig. 3c), implying that the early-flowering phenotype of dd-2 also reflects the down-regulation of MAF4 and MAF5. To address this, we confirmed the levels of MAF4 and MAF5 expression among Col-0, flc-3, dd-2, and dd-2 flc-3 triple mutants by qRT-PCR (Fig. 3d and 3e). Consistently, the expression levels of both MAF4 and MAF5 are lower in dd-2 mutants but not affected by flc-3 mutations, partially explaining earlier flowering of dd-2 mutants (Fig. 3c). Consistent with flowering behaviors, expression of a floral integrator, FT, was increased in dd-2, flc-3, and dd-2;flc-3 triple mutants (Fig. 3f).

DEK3 and DEK4 associate with FLC, MAF4, and MAF5 chromatin

To examine whether DEK3 and DEK4 directly control the transcription of FLC, MAF4, and MAF5, we performed chromatin immunoprecipitation (ChIP) assays to determine the enrichment of DEK3 and DEK4 at these loci using the epitope-tagged complementation lines, gDEK3-cMYC and gDEK4-cMYC (Fig. 1e and 1f). We observed the enrichment of DEK3 throughout the FLC region, including the promoter region (Fig. 4a and 4b). Consistent with the levels of transgene expression (Fig. 1e and Fig. S2b), the enrichment of DEK3 is higher in gDEK3#6 than in gDEK3#11. Similar enrichments were also observed at MAF4 and MAF5 chromatin (Fig. 4c). The enrichment of DEK4 at FLC is also similar to that of DEK3 (Fig. 4d), further confirming partially redundant roles of DEK proteins in the regulation of FLC. Taken together, our results indicated that DEK3 and DEK4 directly associate with FLC, MAF4, and MAF5 chromatin.

Fig. 4.

Fig. 4

Open in a new tab

DEK3 and DEK4 directly binds to FLC, MAF4, and MAF5 chromatin in Arabidopsis. (a) A schematic diagram showing the genome regions of FLC and MAF4/5. Exons are represented by black boxes, while black lines represent introns. DNA fragments amplified in ChIP assays are labeled beneath the genomic regions. (b) Analysis of DEK3 binding to FLC genomic regions. FLC fragments were immunoprecipitated by anti-cMYC were quantified by qPCR and normalized to the Input. Relative levels in gDEK3#6 and gDEK3#11 over dek3-2 (Control) are presented, Error bars: ± s.d. (n = 2). (c) Analysis of DEK3 binding to MAF4 and MAF5 genomic regions. (d) Analysis of DEK4 binding to FLC genomic regions. Relative levels in gDEK4#21 over dd-2 (Control) are presented. Error bars: ± s.d. (n = 2). (e) DEK3 and DEK4 interact with Histone H3 in vivo. Total protein extracts from 35S::DEK3-GFP, 35S::DEK4-GFP, and 35S-GFP lines were subjected to IP with GFP-trap Dynabeads, and immunoprecipitated proteins were detected by western blot by using the antibodies against Histone H3 and GFP. The input represents 5% of the amount of proteins used for IP. The predicated molecular mass of DEK3-GFP and DEK4-GFP was ~110 kDa (*), but mainly migrates on SDS-PAGE with an apparent size of ~200 kDa (**). This is consistent with a previous report from Drosophila DEK protein, suggesting extensive protein modifications (Sawatsubashi et al., 2010).

It has been shown that DEK3 directly interacts with Histone H3 and H4 (Waidmann et al., 2014). To validate the interaction between DEK3 and Histone H3 in vivo by co-immunoprecipitation (Co-IP) assays, we first generated the GFP-tagged DEK3 (35S-DEK3-GFP) and DEK4 lines (35S-DEK4-GFP) in Arabidopsis (Fig. S7a). Consistent with previous studies (Waidmann et al., 2014; Anna Brestovitsky, 2019), both DEK3-GFP and DEK4-GFP were mainly localized in the nuclei and cause late flowering in Arabidopsis (Fig. S7b and 7c). Then the GFP-trap Dynabeads was used to purify DEK3-GFP-containing complex from total protein extracts. As shown in Fig. 4e, Histone H3 was co-precipitated with DEK3 in vivo. Like DEK3, DEK4-GFP was also able to precipitate Histone H3 in vivo (Fig. 4e), confirming that both DEK3 and DEK4 closely interact with Histone octamer in vivo.

DEK3 and DEK4 affect histone modifications at FLC, MAF4, and MAF5

Direct associations of DEK proteins with FLC, MAF4, and MAF5 chromatin and with Histone octamer in vivo prompted us to investigate the effect of DEK proteins on histone modifications at target loci. FLC chromatin has been an excellent model system to study a number of histone modifications and chromatin-modifying enzymes in Arabidopsis and functional relevance of several histone modifications has been well established (Whittaker & Dean, 2017). Interestingly, the level of trimethylation at Histone H3 Lys 4 (H3K4me3), a representative histone mark for gene activation, is not altered at the FLC in dd-2 mutants compared to the wild-type Col-0 (Fig. 5a), despite the difference in the level of transcription. Trimethylation at Histone H3 Lys 27 (H3K27me3) and trimethylation at Histone H3 Lys 36 (H3K36me3) are two histone marks that antagonize each other to fine tune FLC expression (Kim et al., 2009; He, 2012). Indeed, the level of H3K27me3 enrichment is higher, whereas the level of H3K36me3 enrichment is lower at FLC in dd-2 mutants compared to the wild-type Col-0 (Fig. 5b and 5c).

Fig. 5.

Fig. 5

Open in a new tab

DEK3 and DEK4 affect histone modifications at FLC, MAF4, and MAF5 loci in Arabidopsis. (a) to (c) ChIP-qPCR analysis of H3K4me3 (a), H3K27me3 (b), and H3K26me3 (c) levels at FLC in the wild-type (Col-0) and dd-2 seedlings. (d) to (f). ChIP-qPCR analysis of H3K4me3 (d), H3K27me3 (e), and H3K36me3 (f) levels at MAF4 and MAF5 loci in the wild-type (Col-0) and dd-2 seedlings. Each examined region was normalized to the AGAMOUS (AG). Error bars: ± s.d. (n = 2). Asterisks indicate statistically significant differences in the indicated plant pairs (two-tailed paired Student’s t-test, p<0.01).

Similarly, the levels of H3K4me3 enrichment are not altered at MAF4 and MAF5 chromatin in dd-2 mutants (Fig. 5d). Consistent with the down regulation of MAF4 and MAF5 expression, two repressive marks, H3K27me3 levels were significantly increased in dd-2 mutants, whereas an activation mark H3K36me3 levels was decreased in dd-2 mutants (Fig 5e and 5f). Unlike MAF4 and MAF5, the expressions of MAF1~3 are not affected in dek mutants (Fig. 2e), despite MAF2~5 being clustered together in the Arabidopsis genome. Consistent with that, the enrichment level of H3K27me3 and H3K36me3 at these loci was not affected in dd-2 mutants (Fig. S8). Taken together, our data suggest that DEK3 and DEK4 affect the levels of H3K27me3 and H3K36me3 at FLC, MAF4, and MAF5 chromatin through the direct association.

DEK3 and DEK4 increase the enrichment of RNA polymerase II at FLC, MAF4, and MAF5 loci

A previous study showed that DEK3 interacts with Topoisomerase Iα (TOP1α) in vivo (Waidmann et al., 2014). In addition, loss of function mutations of TOP1α result in early flowering (Gong et al., 2017). TOP1α promotes the association of RNA Polymerase II complexes (Pol II) with FLC, MAF4, and MAF5 chromatin and thus further promotes their transcription (Gong et al., 2017). Therefore, we test whether DEK3 and DEK4 function through TOP1α to facilitate the transcription of FLC, MAF4, and MAF5. We first introduced dek3-2 into top1α to create the double mutants by genetic crossing (Fig. 6a). Because DEK4 is closely linked to TOP1α locus, we could not generate the double mutants between dek4-2 and top1α. Nevertheless, both top1α alleles are completely epistatic to dek3-2 mutants, resulting in earlier flowering than dek3-2 mutants but not earlier than top1α mutants (Fig. 6a), indicating that DEK3 and TOP1α function together.

Fig. 6.

Fig. 6

Open in a new tab

DEK3 and DEK4 directly interact with Pol II and promote the association of Pol II to FLC, MAF4, and MAF5 loci in Arabidopsis. (a) Flowering times of indicated genotypes in LD. Flowering time was measured by the number of rosette leaves at bolting. Error bars: ± s.d. (n = 12). One-way ANOVA with Tukey’s multiple comparison test was conducted; asterisks indicate p< 0.01 of distinct groups. (b) and (c) ChIP-qPCR analysis of the association of Pol II (ser5P) to FLC (b) and MAF4 and MAF5 (c) loci in the wild-type (Col-0) and dd-2 seedlings. Each examined region was normalized to the Input, Error bars: ± s.d. (n = 2). Asterisks indicate statistically significant differences in the indicated plant pairs (two-tailed paired Student’s t-test, p<0.01). (d) DEK3 and DEK4 interact with Pol II in vivo. Total protein extracts from 35S::DEK3-GFP expressing lines or 35S-GFP lines were subjected to IP with GFP-trap Dynabeads, and immunoprecipitated proteins were detected by western blot by using the antibody against Pol IIser5P. The input represents 5% of the amount of proteins used for IP.

This led us to test whether DEK3 and DEK4 also affect the accessibility of Pol II to FLC, MAF4, and MAF5 loci. Indeed, the association of Pol II with FLC, MAF4, and MAF5 was significantly decreased in dd-2 mutants (Fig. 6b and 6c). In addition, Co-IP assays using the GFP-tagged DEK3 (35S-DEK3-GFP) and DEK4 (35S-DEK4-GFP) transgenic lines revealed that both DEK3 and DEK4 physically associated with Pol II in vivo (Fig. 6d), indicating that DEK proteins associate with the basal transcription machinery. To test whether DEK3 and DEK4 affect nucleosome positioning, we further investigate the nucleosome occupancy levels in dd-2 mutants at FLC, MAF4 and MAF5 loci. Consistent with previous report that the top1α mutation did not change the nucleosome occupancy (Gong et al., 2017), we also did not observe any significant change in nucleosome occupancy between dd-2 and wild type (Col-0) (Fig. S9). Taken together, DEK3 and DEK4 function through TOP1α in the recruitment of the transcriptional initiation complex to these floral repressor loci during the floral transition.

Discussion

DEK domain-containing proteins are evolutionally conserved and have been implicated in several cellular processes in many eukaryotes (Sanden & Gullberg, 2015). DEK has no known enzymatic activity and instead function as an architectural chromatin protein that binds to DNA, chromatin, and histones (Sanden & Gullberg, 2015; de Albuquerque Oliveira et al., 2018). There are four DEK homologs in Arabidopsis, but their biological roles in plant development remain largely elusive. In this report, we showed that DEK3 and DEK4 redundantly control flowering time in Arabidopsis. Transcriptome and genetic analysis showed that DEK3 and DEK4 are necessary for the active transcription of three related floral repressors, FLC, MAF4, and MAF5. ChIP analysis showed that DEK3 and DEK4 directly associate with FLC, MAF4, and MAF5 chromatin. Interestingly, DEK3 and DEK4 proteins are distributed rather evenly throughout the target loci, which may reflect their strong association with core histone proteins. Previous reports also reported broad distributions of DEK3 at its target loci (Waidmann et al., 2014; Anna Brestovitsky, 2019).

DEK plays a dual role in the transcriptional regulation of its target genes (Sanden et al., 2014; Anna Brestovitsky, 2019). In Drosophila, DEK coactivates the ecdysone nuclear receptor by acting as a histone chaperone. DEK preferentially associates with activating histone modifications, including H3K4me3 (Sawatsubashi et al., 2010). In addition, DEK physically associates with Pol II to promote transcription (Sawatsubashi et al., 2010). On the other hand, DEK has also been shown to function to repress transcription by inhibiting histone acetyltransferases, p300, and PCAF (Ko et al., 2006).

Our results suggested that DEK4 and DEK5 act as activators for FLC, MAF4, and MAF5 by creating a chromatin structure that favors the transcriptional activation. Two antagonistic histone modifications, H3K27me3 and H3K36me3, were affected at FLC, MAF4, and MAF5 loci by the mutations in DEK3 and DEK4. In dd-2 double mutants, the level of H3K36me3 decreased, whereas the level of H3K27me3 increased at these loci. Our results also showed that both DEK3 and DEK4 associate with Histone proteins in vivo (Fig. 4e). Taken together, DEK3 and DEK4 function as chromatin architectural proteins that modulate histone modifications at their target loci, including FLC.

It is interesting to note that top1α mutants are completely epistatic to dek3 mutants in terms of flowering time, implicating that DEK3 and DEK4 function together with TOP1α to activate FLC, MAF4, and MAF5. TOP1α is necessary for effective Pol II recruitment (Gong et al., 2017). Similarly, we showed that DEK3 and DEK4 are required to recruit Pol II effectively (Fig. 6). Our work showed that DEK proteins play roles in the floral transition by modulating chromatin states to be active chromatin at floral repressor loci. Therefore, evolutionarily conserved DEK proteins function as chromatin architectural modulators to control multiple cellular processes in eukaryotes.

Supplementary Material

Supplementary Data

Table S1. Primers used in this study.

Fig. S1. DEK1 and DEK2 do not control flowering time in Arabidopsis.

Fig. S2. Immunoblotting detects the DEK3 and DEK4 in complementation lines.

Fig. S3. DEK3 and DEK4 regulate the expression of FLC, MAF4 and MAF5.

Fig. S4. DEK3 and DEK4 regulate the expression of FLC under short-day conditions.

Fig. S5. FLC expression patterns in dd-2 and Col-0.

Fig. S6. DEK3 and DEK4 were not involved in vernalization response in Arabidopsis.

Fig. S7. DEK3 and DEK4 over-expression delayed flowering in Arabidopsis.

Fig. S8. DEK3 and DEK4 do not affect histone modifications at MAF1, MAF2 and MAF3 loci.

Fig. S9. DEK3 and DEK4 not affect nucleosome occupancy of the FLC, MAF4, and MAF5 loci.

Supplementary Table S2

Table S2. DEGs in dek mutants.

Summary.

  • Evolutionarily conserved DEK domain-containing proteins have been implicated in multiple chromatin-related processes, mRNA splicing, and transcriptional regulation in eukaryotes.

  • Here, we show that two DEK proteins, DEK3 and DEK4, control the floral transition in Arabidopsis. DEK3 and DEK4 directly associate with chromatin of related flowering repressors, FLOWERING LOCUS C (FLC), and its two homologs, MADS AFFECTING FLOWERING4 (MAF4) and MAF5, to promote their expression.

  • The binding of DEK3 and DEK4 to histone octamer in vivo affects histone modifications at FLC, MAF4 and MAF5 loci. In addition, DEK3 and DEK4 interact with RNA polymerase II and promote the association of RNA polymerase II with FLC, MAF4 and MAF5 chromatin to promote their expression.

  • Our results show that DEK3 and DEK4 directly interact with chromatin to facilitate the transcription of key flowering repressors and thus prevent precocious flowering in Arabidopsis.

Acknowledgments

We wish to thank Maddie Brightbill for comments on the manuscript. This work was supported by NIH R01GM100108 and NSF IOS 1656764 to S. S.

Data Availability

The RNA-Seq datasets generated from this study have been deposited in the Gene Expression Omnibus (GEO) under accession GSE163742.

Reference

  1. Alexiadis V, Waldmann T, Andersen J, Mann M, Knippers R, Gruss C. 2000. The protein encoded by the proto-oncogene DEK changes the topology of chromatin and reduces the efficiency of DNA replication in a chromatin-specific manner. Genes Dev 14(11): 1308–1312. [PMC free article] [PubMed] [Google Scholar]
  2. Anna Brestovitsky DE, Sascha Waidmann, Maslen Sarah L., Balcerowicz Martin, Cortijo Sandra, Charoensawan Varodom, Martinho Claudia, Rhodes Daniela, Jonak Claudia, Wigge Philip A. 2019. DEK influences the trade-off between growth and arrest via H2A.Z-nucleosomes in Arabidopsis. bioRxiv. [Google Scholar]
  3. Campillos M, Garcia MA, Valdivieso F, Vazquez J. 2003. Transcriptional activation by AP-2alpha is modulated by the oncogene DEK. Nucleic Acids Res 31(5): 1571–1575. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Clough SJ, Bent AF. 1998. Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana. Plant J 16(6): 735–743. [DOI] [PubMed] [Google Scholar]
  5. de Albuquerque Oliveira AC, Kappes F, Martins DBG, de Lima Filho JL. 2018. The unique DEK oncoprotein in women’s health: A potential novel biomarker. Biomed Pharmacother 106: 142–148. [DOI] [PubMed] [Google Scholar]
  6. Gong X, Shen L, Peng YZ, Gan Y, Yu H. 2017. DNA Topoisomerase Ialpha Affects the Floral Transition. Plant Physiol 173(1): 642–654. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. He Y 2012. Chromatin regulation of flowering. Trends Plant Sci 17(9): 556–562. [DOI] [PubMed] [Google Scholar]
  8. He Y, Michaels SD, Amasino RM. 2003. Regulation of flowering time by histone acetylation in Arabidopsis. Science 302(5651): 1751–1754. [DOI] [PubMed] [Google Scholar]
  9. Ivanauskiene K, Delbarre E, McGhie JD, Kuntziger T, Wong LH, Collas P. 2014. The PML-associated protein DEK regulates the balance of H3.3 loading on chromatin and is important for telomere integrity. Genome Res 24(10): 1584–1594. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Kappes F, Waldmann T, Mathew V, Yu JD, Zhang L, Khodadoust MS, Chinnaiyan AM, Luger K, Erhardt S, Schneider R, et al. 2011. The DEK oncoprotein is a Su(var) that is essential to heterochromatin integrity. Genes & Development 25(7): 673–678. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Kavanaugh GM, Wise-Draper TM, Morreale RJ, Morrison MA, Gole B, Schwemberger S, Tichy ED, Lu L, Babcock GF, Wells JM, et al. 2011. The human DEK oncogene regulates DNA damage response signaling and repair. Nucleic Acids Res 39(17): 7465–7476. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Kim D, Langmead B, Salzberg SL. 2015. HISAT: a fast spliced aligner with low memory requirements. Nat Methods 12(4): 357–360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kim DH, Doyle MR, Sung S, Amasino RM. 2009. Vernalization: winter and the timing of flowering in plants. Annu Rev Cell Dev Biol 25: 277–299. [DOI] [PubMed] [Google Scholar]
  14. Kim DH, Sung S. 2014. Genetic and epigenetic mechanisms underlying vernalization. Arabidopsis Book 12: e0171. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ko SI, Lee IS, Kim JY, Kim SM, Kim DW, Lee KS, Woo KM, Baek JH, Choo JK, Seo SB. 2006. Regulation of histone acetyltransferase activity of p300 and PCAF by proto-oncogene protein DEK. FEBS Lett 580(13): 3217–3222. [DOI] [PubMed] [Google Scholar]
  16. Koleva RI, Ficarro SB, Radomska HS, Carrasco-Alfonso MJ, Alberta JA, Webber JT, Luckey CJ, Marcucci G, Tenen DG, Marto JA. 2012. C/EBPalpha and DEK coordinately regulate myeloid differentiation. Blood 119(21): 4878–4888. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Lee I, Michaels SD, Masshardt AS, Amasino RM. 1994. The Late-Flowering Phenotype of Frigida and Mutations in Luminidependens Is Suppressed in the Landsberg Erecta Strain of Arabidopsis. Plant Journal 6(6): 903–909. [Google Scholar]
  18. Li Z, Jiang D, He Y. 2018. FRIGIDA establishes a local chromosomal environment for FLOWERING LOCUS C mRNA production. Nat Plants 4(10): 836–846. [DOI] [PubMed] [Google Scholar]
  19. Luo X, He Y. 2020. Experiencing winter for spring flowering: A molecular epigenetic perspective on vernalization. J Integr Plant Biol 62(1): 104–117. [DOI] [PubMed] [Google Scholar]
  20. Michaels SD, Amasino RM. 1999. FLOWERING LOCUS C encodes a novel MADS domain protein that acts as a repressor of flowering. Plant Cell 11(5): 949–956. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Michaels SD, Himelblau E, Kim SY, Schomburg FM, Amasino RM. 2005. Integration of flowering signals in winter-annual Arabidopsis. Plant Physiol 137(1): 149–156. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Qian S, Lv X, Scheid RN, Lu L, Yang Z, Chen W, Liu R, Boersma MD, Denu JM, Zhong X, et al. 2018. Dual recognition of H3K4me3 and H3K27me3 by a plant histone reader SHL. Nat Commun 9(1): 2425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Sanden C, Gullberg U. 2015. The DEK oncoprotein and its emerging roles in gene regulation. Leukemia 29(8): 1632–1636. [DOI] [PubMed] [Google Scholar]
  24. Sanden C, Jarvstrat L, Lennartsson A, Brattas PL, Nilsson B, Gullberg U. 2014. The DEK oncoprotein binds to highly and ubiquitously expressed genes with a dual role in their transcriptional regulation. Mol Cancer 13: 215. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Sawatsubashi S, Murata T, Lim J, Fujiki R, Ito S, Suzuki E, Tanabe M, Zhao Y, Kimura S, Fujiyama S, et al. 2010. A histone chaperone, DEK, transcriptionally coactivates a nuclear receptor. Genes Dev 24(2): 159–170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Shibata T, Kokubu A, Miyamoto M, Hosoda F, Gotoh M, Tsuta K, Asamura H, Matsuno Y, Kondo T, Imoto I, et al. 2010. DEK oncoprotein regulates transcriptional modifiers and sustains tumor initiation activity in high-grade neuroendocrine carcinoma of the lung. Oncogene 29(33): 4671–4681. [DOI] [PubMed] [Google Scholar]
  27. Soares LM, Zanier K, Mackereth C, Sattler M, Valcarcel J. 2006. Intron removal requires proofreading of U2AF/3’ splice site recognition by DEK. Science 312(5782): 1961–1965. [DOI] [PubMed] [Google Scholar]
  28. Sung S, Amasino RM. 2004. Vernalization in Arabidopsis thaliana is mediated by the PHD finger protein VIN3. Nature 427(6970): 159–164. [DOI] [PubMed] [Google Scholar]
  29. von Lindern M, Fornerod M, van Baal S, Jaegle M, de Wit T, Buijs A, Grosveld G. 1992. The translocation (6;9), associated with a specific subtype of acute myeloid leukemia, results in the fusion of two genes, dek and can, and the expression of a chimeric, leukemia-specific dek-can mRNA. Mol Cell Biol 12(4): 1687–1697. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Waidmann S, Kusenda B, Mayerhofer J, Mechtler K, Jonak C. 2014. A DEK domain-containing protein modulates chromatin structure and function in Arabidopsis. Plant Cell 26(11): 4328–4344. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Whittaker C, Dean C. 2017. The FLC Locus: A Platform for Discoveries in Epigenetics and Adaptation. Annu Rev Cell Dev Biol 33: 555–575. [DOI] [PubMed] [Google Scholar]
  32. Yoo SK, Chung KS, Kim J, Lee JH, Hong SM, Yoo SJ, Yoo SY, Lee JS, Ahn JH. 2005. CONSTANS activates SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1 through FLOWERING LOCUS T to promote flowering in Arabidopsis. Plant Physiol 139(2): 770–778. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Data

Table S1. Primers used in this study.

Fig. S1. DEK1 and DEK2 do not control flowering time in Arabidopsis.

Fig. S2. Immunoblotting detects the DEK3 and DEK4 in complementation lines.

Fig. S3. DEK3 and DEK4 regulate the expression of FLC, MAF4 and MAF5.

Fig. S4. DEK3 and DEK4 regulate the expression of FLC under short-day conditions.

Fig. S5. FLC expression patterns in dd-2 and Col-0.

Fig. S6. DEK3 and DEK4 were not involved in vernalization response in Arabidopsis.

Fig. S7. DEK3 and DEK4 over-expression delayed flowering in Arabidopsis.

Fig. S8. DEK3 and DEK4 do not affect histone modifications at MAF1, MAF2 and MAF3 loci.

Fig. S9. DEK3 and DEK4 not affect nucleosome occupancy of the FLC, MAF4, and MAF5 loci.

NIHMS1792938-supplement-Supplementary_Data.pdf (2MB, pdf)
Supplementary Table S2

Table S2. DEGs in dek mutants.

NIHMS1792938-supplement-Supplementary_Table_S2.xlsx (66.5KB, xlsx)

Data Availability Statement

The RNA-Seq datasets generated from this study have been deposited in the Gene Expression Omnibus (GEO) under accession GSE163742.

RESOURCES