The Arabidopsis thaliana ecotype Col-0 and mutants pwr-2 (SALK_071811), pwr-3 (SALK_006823), hda9-1 (GK_305G03), hda9-2 (SALK_007123), and hos15-2 (GK_785B10) mutants, and HOS15pro::HOS15-HA/hos15-2 (Park et al., 2018a), HOS15pro::HOS15-FLAG/hos15-2 (Park et al., 2018b) transgenic plants were used in this study. Double (pwr-2 hos15-2, hda9-2 hos15-2, pwr-2 hda9-2) and triple mutants (pwr-2 hda9-2 hos15-2) were generated by crossing. Seeds were surface-sterilized in a solution containing 3% (v/v) sodium hypochlorite solution (Yakuri Pure Chemicals, Kyoto, Japan) for 5 min and then rinsed five times with sterilized water. After stratification at 4°C for 3 days in the dark, the plants were grown on half-strength Murashinge and Skoog (MS) medium or soil at 23°C ± 1°C under 16-h light/8-h-dark conditions with 100 µM m^-2s^-1 light intensity.

Flowering time analysis of the WT and mutant plants was performed at 23°C ± 1°C in long-day conditions. Flowering time was measured by counting the number of days-to-flowering and the total number of rosette leaves from the primary meristems at bolting.

Total RNA was extracted from two-week-old plants using the RNeasy Plant Mini Kit (Qiagen) and treated with DNase (Sigma-Aldrich). cDNA was synthesized using reverse transcriptase (Toyobo, Osaka, Japan). The SYBR Green PCR Master Mix Kit (Bio-Rad) was used for quantitative PCR using the indicated gene-specific primers. Primer sequences are provided in Supplementary Table 1 . Relative expression levels were analyzed using the comparative cycle threshold method (Bio-Rad).

Total protein was isolated from Arabidopsis seedlings using protein extraction buffer containing 150 mM NaCl, 100 mM Tris-Cl (pH 7.5), 0.5% (v/v) NP-40, 1 mM EDTA, 3 mM DTT, and protease inhibitors (1 mM PMSF, 5 μg mL^-1 leupeptin, 1 μg mL^-1 aprotinin, 1 μg mL^-1 pepstatin, 5 μg mL^-1 antipain, 5 μg mL^-1 chymostatin, 2 mM Na₂VO₃, 2 mM NaF, and 50 mM MG132). The proteins were separated and detected using SDS-PAGE and immunoblottied with the appropriate antibodies, anti-HOS15 (Park et al., 2018a; Lim et al., 2020), anti-α-tubulin (Sigma-Aldrich), anti-Ac-H3 (Millipore), anti-H3 (Abcam), and anti-FLAG (Sigma-Aldrich) for 2 h at room temperature or overnight at 4°C. Membranes were developed using peroxidase-conjugated secondary anti-rabbit IgG (Santa Cruz Biotechnology) and anti-mouse IgG (Santa Cruz Biotechnology). Western blot membranes were incubated with Clarity™ Western ECL Substrate solution (Bio-Rad), and signals were detected using an imaging system (ChemiDoc™MP, Bio-Rad).

Nuclei were extracted from two-week-old plants using Honda’s buffer (0.4 M sucrose, 2.5% Ficoll 400, 5% dextran T-40, 10 mM MgCl₂, 25 mM Tris-Cl (pH 7.5), 10 mM β-mercaptoethanol, 100 mg/mL phenylmethylsulfonyl fluoride, 0.5 mg/mL antipain, and 0.5 mg/mL leupeptin) (Folta and Kaufman, 2006; Park et al., 2018b). The samples were filtered through a 60-µm nylon filter (Millipore) and incubated on ice for 15 min. Triton X-100 was added at a final concentration of 0.5% and the sample was centrifuged at 1500g for 5 min. The supernatants were transferred to new tubes, and cytosolic proteins were used. The pellet was washed with 1 mL Honda’s buffer containing 0.1% Triton X-100. The cells were gently resuspended and centrifuged at 100g for 5 min to pellet the starch and cell debris. The supernatant was transferred to a tube and centrifuged at 1800g for 5 min to pellet the nuclei. The nuclear and cytosolic extracts were suspended in SDS-PAGE loading buffer. Immunoblots were performed using anti-HOS15, anti-α-tubulin (Sigma), and anti-H3 (Abcam) antibodies, and antigen proteins were visualized by chemiluminescence using ECL-detecting reagent (Bio-Rad).

The chromatin immunoprecipitation (ChIP) assay was performed as described by Lim et al. (2020). Nuclear proteins were extracted from two-week-old Arabidopsis plants, and immunoprecipitation was performed using salmon sperm DNA/protein A agarose beads (Upstate Biotechnology) fused to anti-HOS15 and anti-acetylated H3 antibodies (Millipore). Beads treated with anti-rabbit IgG were used as negative controls for each ChIP assay. The immunoprecipitated DNA was purified by phenol-chloroform-isoamyl (PCI) extraction and ethanol precipitation. The precipitated DNA was dissolved in TE buffer and quantified using RT-qPCR. The primers used in the ChIP assays are listed in Supplementary Table 1 , and an UBQ10 DNA fragment was used for normalization.

To investigate the function of the HOS15-PWR-HDA9 complex in regulating the flowering time, we analyzed flowering phenotypes of the loss-of-function hos15-2 mutant and three complementation lines expressing the construct HOS15pro::HOS15-HA in the hos15-2 background (Park et al., 2018a). As expected, hos15-2 mutant plants exhibited early flowering (Park et al., 2019), while the flowering time of complementation lines was similar to that of WT plants ( Supplemental Figures 1A–C ). Using RNA-seq analysis data, we previously showed that the expression level of the AGL19 was upregulated in hos15-2 mutants (Park et al., 2019), which could promote flowering time (Kang et al., 2015; Park et al., 2019). To confirm this, we analyzed the transcript level of AGL19 using RT-qPCR, which was significantly increased in the hos15-2 mutant. In contrast, the complementation lines possessed a similar level of AGL19 transcription to that of the WT plants ( Supplemental Figure 1D ; Supplemental Figure 2 ). These results suggest that the early flowering phenotype of hos15-2 might be due to the upregulation of AGL19.

The transcript level of AGL19 has previously been reported to be upregulated in pwr and hda9 knockout mutants, and as a result, these mutants show early flowering phenotypes (Kang et al., 2015; Kim et al., 2016). Since HOS15 knockout plants also show early flowering and the upregulation of AGL19 transcript levels ( Supplemental Figure 1 ), we assumed that HOS15 might work together with PWR-HDA9 in the same repressor complex, thereby negatively regulating flowering time by repressing AGL19 transcription. To test our hypothesis, we generated double and triple (pwr hos15, hda9 pwr, hda9 hos15, hda9 hos15 pwr) mutant plants and tested their flowering phenotypes. As expected, double and triple knockout mutants of the HOS15-PWR-HAD9 complex showed an early flowering phenotype statistically similar to single mutants of HOS15, PWR, and HDA9, signifying that these mutations were not additive ( Figure 1A–C ). In addition to flowering phenotypes, the transcript level of AGL19 was upregulated in single, double, and triple mutant plants compared to that in WT plants ( Figure 1D ). Furthermore, double and triple mutants of HOS15, PWR, and HDA9 exhibited dwarf phenotypes, such as small plant size and short and blunt-end siliques, which are common phenotypes of hos15-2, pwr, and hda9 single mutants ( Supplemental Figures 3 and 4 ) (Kang et al., 2015; Kim et al., 2016; Mayer et al., 2019; Lim et al., 2020). Overall, these findings indicate that the HOS15-PWR-HDA9-complex negatively regulates flowering time by inhibiting AGL19 expression.

HDA9 negatively regulates AGL19 expression by associating with the promoter region of AGL19 and promoting histone deacetylation (Kang et al., 2015). Therefore, we tested whether HOS15 is associated with the promoter region of AGL19 to modulate chromatin status. A chromatin immunoprecipitation (ChIP) assay showed that HOS15 was indeed associated with the E region of the AGL19 promoter, in a region already known for HDA9 association ( Figure 2 ; Kang et al., 2015). These data suggest that HOS15 binds to the promoter of AGL19 to co-regulate the histone acetylation status of AGL19 chromatin.

HOS15 forms a complex with PWR and HDA9 (Park et al., 2018b; Suzuki et al., 2018; Mayer et al., 2019). Furthermore, HOS15, PWR, and HDA9 were predominantly localized in the nucleus (Zhu et al., 2008; Kang et al., 2015; Lim et al., 2020), and HDA9 accumulation was reduced in the nuclei of pwr and hos15 mutant plants (Chen et al., 2016; Mayer et al., 2019). Based on these precedents, we investigated whether the accumulation of HOS15 in nuclei also required PWR-HDA9. A nuclear-cytoplasm fractionation analysis using pwr and hda9 mutant plants showed that HOS15 nuclear accumulation was greatly decreased in pwr and hda9 mutants as compared with WT ( Figure 3 ). Notably, compared with WT plants, HOS15 transcript levels were not changed in pwr and hda9 mutants; in contrast, the total protein amount was decreased in these mutants ( Supplemental Figure 5 ). HOS15 also interacts with other RPD3-type class 1 histone deacetylases, such as HDA6 and HDA19 (Park et al., 2018a; Park et al., 2018b). However, unlike hda9, hda6, and hda19 mutants did not exhibit early flowering; instead, they exhibited partial late-flowering phenotypes ( Supplemental Figures 6A, B ). Furthermore, HOS15 total protein remained unchanged in hda6 and hda19 mutant plants compared with that in WT plants ( Supplemental Figure 6C ). Overall, these results suggest that HOS15 forms a complex with PWR-HDA9 in the nucleus and regulates floral transition by repressing AGL19 expression.

Previous reports have revealed that HOS15, PWR, and HDA9 repress target gene expression by regulating histone H3 acetylation status (Kim et al., 2016; Mayer et al., 2019; Park et al., 2019). To investigate whether HOS15 interacts with H3, we performed a co-IP assay using two independent HOS15pro::HOS15-Flag/hos15-2 complementation plant lines (Park et al., 2018b). As shown in Figure 4A , HOS15 interacted with H3 in vivo. As HOS15 stability in the nucleus is dependent on PWR-HDA9, we tested HOS15 and H3 interactions in pwr and hda9 mutants. Notably, the interaction of HOS15 with H3 was drastically reduced in pwr and hda9 mutants compared with that in the WT plants ( Figure 4B ). In addition, total histone acetylation levels were increased in hos15, pwr and hda9 single, double, and triple mutant plants compared with WT plants, and the increment was similar between single and higher order mutants ( Figure 5A ) (Zareen et al., 2022). In line with this, we next tested the acetylation status of AGL19 chromatin in hos15, pwr and hda9 single and hos15/pwr/hda9 triple mutant plants using a ChIP assay with acetylated H3 (AcH3) antibody. H3 acetylation of AGL19 chromatin was higher in hos15-2 mutant plants than in the WT plants ( Figure 5B , Supplemental Figure 7 ). Consistent with the notion that HOS15 functions together with the PWR-HDA9 complex, no significant difference in the increase in H3 acetylation level was observed between these single and triple mutant plants ( Figure 5B ). Taken together, these findings suggest that HOS15, PWR, and HDA9 work together in the same repressor complex and deacetylate AGL19 chromatin region, thereby negatively regulating AGL19 expression.

Epigenetic regulation plays an essential role in various signaling pathways in plants. Our study revealed the role of the HOS15-PWR-HDA9 co-repressor complex in the epigenetic regulation of flowering time through the expression of AGL19, a flowering-regulating gene (Kim et al., 2013; Kang et al., 2015). Although the regulatory pathways leading to flowering in Arabidopsis have been well documented, there is still room for further investigation to uncover the complete mechanisms of floral transitions. Using genetic interactions, proteomics, and chromatin immunoprecipitation assays, we found that the HOS15-PWR-HDA9 co-repressor complex inhibits AGL19 expression through histone modification of AGL19 chromatin thereby negatively regulating flowering time in Arabidopsis ( Figure 6 ). Loss-of-function single, double, or triple mutants of the HOS15-HDA9-PWR complex resulted in early flowering phenotypes ( Figure 1 ). Furthermore, the expression of AGL19 was strongly upregulated in hos15, hda9 and pwr mutants, which highlights the involvement of HOS15-HDA9-PWR complex in floral transition. The diurnal expression pattern of AGL19 was altered in the triple and the double mutants hos15 hda9 and pwr hda9 with a decline at ZT12 ( Figure 1 ). The reason for this behavior is unknown, but the HOS15/PWR/HDA9 complex is likely to impact the expression of other circadian and photoperiodic genes whose altered expression may, in turn, impinge on the diurnal rhythm of AGL19 transcripts enhancing the oscillations (Brambilla and Fornara, 2017).

Previously we have shown that HDA9 and HOS15 regulate flowering time by interfering with GI transcription (Park et al., 2019). We have also shown that PWR is required for HOS15-mediated HD2C degradation upon cold stress (Lim et al., 2020). These findings suggest that HDA9 and PWR are required for functional HOS15. To confirm this, we analyzed HOS15 accumulation in hda9 and pwr mutants both in nucleus and cytosol. As shown in Figure 3 , HOS15 accumulation in the nucleus was greatly reduced in hda9 and pwr mutants compared with WT. By contrast, HOS15 accumulation in the cytosol was almost similar among all tested lines ( Figure 3 ), suggesting that HDA9 and PWR are required for nuclear accumulation of HOS15. Furthermore, HOS15 interaction with H3 was impaired in hda9 and pwr mutants ( Figure 4 ) which further strengthened the notion that functional HDA9 and PWR are critical for HOS15’s nuclear activity.

Recent reports have shown that HOS15-HDA9-PWR is a co-repressor complex which modulates histone acetylation status thereby regulating physiological processes including senescence (Mayer et al., 2019; Zareen et al., 2022). Consistent with previous report, acetylated H3 was accumulated abundantly in hos15, hda9 and pwr single, double, and triple mutants. In addition to this, we have also found that HOS15-HDA9-PWR complex regulates flowering time through regulation of AGL19 chromatin ( Figure 5 ). Using chromatin immunoprecipitation assay, we found that AGL19 chromatin was hyperacetylated in loss of function mutants of HOS15-HDA9-PWR complex ( Figure 5 ), highlighting the importance of this complex in floral transition via AGL19 regulation. In the WT plants, the HOS15-PWR-HDA9 co-repressor complex negatively regulated the expression of AGL19 by promoting histone deacetylation at the AGL19 promoter, which compacts chromatin. In the absence of PWR and HDA9, the HOS15 protein was unstable in the nucleus. The removal of either one of the components of the repressor complex prevented histone deacetylation, thereby relaxing AGL19 chromatin, the increase in the expression of AGL19 and accelerating the flowering time ( Figure 6 ). This study revealed that HOS15 together with HDA9 and PWR control flowering time by repressing AGL19 through increased histone deacetylation.