AGAMOUS terminates floral stem cell maintenance in Arabidopsis by directly repressing WUSCHEL through recruitment of Polycomb Group proteins

Abstract

Floral stem cells produce a defined number of floral organs before ceasing to be maintained as stem cells. Therefore, floral stem cells offer an ideal model to study the temporal control of stem cell maintenance within a developmental context. AGAMOUS (AG), a MADS domain transcription factor essential for the termination of floral stem cell fate, has long been thought to repress the stem cell maintenance gene WUSCHEL (WUS) indirectly. Here, we uncover a role of Polycomb Group (PcG) genes in the temporally precise repression of WUS expression and termination of floral stem cell fate. We show that AG directly represses WUS expression by binding to the WUS locus and recruiting, directly or indirectly, PcG that methylates histone H3 Lys-27 at WUS. We also show that PcG acts downstream of AG and probably in parallel with the known AG target KNUCKLES to terminate floral stem cell fate. Our studies identify core components of the network governing the temporal program of floral stem cells.

Figure 1.

Phenotypes of ag and clf Single and Double Mutants. (A) A wild-type (Ler) flower. (B) An ag-10 flower with a slightly enlarged gynoecium. (C) A clf-47 flower. (D) An ag-10 clf-47 flower with a much more enlarged gynoecium compared with ag-10. (E) and (F) Longitudinal sections through stage 11 flowers of ag-10 clf-47 (E) and ag-10 (F) genotypes. In (E), the floral meristem continued to generate organs (arrow) inside the carpels. (G) A clf-2 flower. (H) An ag-10 clf-2 flower with similar phenotypes to those of ag-10 clf-47. (I) Siliques from plants of the indicated genotypes. Most siliques on an ag-10 plant were long and thin (represented by the two on the left); one to a few siliques were short and bulged (represented by the one on the right). Most siliques from ag-10 clf-47 or ag-10 clf-2 plants were short and bulged. (J) Siliques from F1 plants of the cross between ag-10 clf-47 and ag-10 clf-2. The siliques were similar in morphology to those of ag-10 clf-47 or ag-10 clf-2. Bars = 1 mm in (A) to (C) and (G) to (I), 0.75 mm in (D), 0.5 mm in (J), and 100 μm in (E) and (F).

Figure 2.

Expression Patterns of WUS and STM in ag-10 and ag-10 clf-47 Flowers. (A) to (D) In situ hybridization with a WUS antisense probe. The arrows indicate WUS signals. (A) and (B) WUS expression was detected in a stage 7 (A) but not a stage 9 (B) ag-10 flower. (C) and (D) WUS expression was detected in ag-10 clf-47 flowers at stage 9 (C) and stage 11 (D). (E) to (G) In situ hybridization using an STM antisense probe. The arrows indicate STM signals. (E) and (F) STM expression was detected at stage 7 (E) but not stage 9 (F) in ag-10 flowers. (G) STM expression was detected in a stage 12 ag-10 clf-47 flower. (H) A longitudinal section of a stages 8 or 9 ag-10 clf-47 pWUS:GUS flower showing GUS staining inside the carpels. (I) and (J) GUS staining in pWUS:GUS (I) and ag-10 clf-47 pWUS:GUS (J) inflorescences. Arrows in (J) indicate GUS signals in the center of stages 8 and 9 flowers. The insets are stages 8 and 9 flowers. The ring-like GUS signals were from anthers. The GUS signals inside the ring in (J) were from the floral meristem. Bars = 50 μm in (A) to (H) and 250 μm in (I) and (J).

Figure 3.

Phenotypes of clf-47 and ag-10 clf-47 in Combination with Mutations in Other Floral Meristem Regulators. (A) A wus-1 flower that lacked a full complement of floral organs. (B) An ag-10 clf-47 wus-1 flower, which was similar to wus-1 with respect to floral meristem determinacy. (C) An stm-2 flower that lacked a full complement of floral organs. (D) An ag-10 clf-47 stm-2 flower, which was similar to stm-2 flowers in terms of floral meristem determinacy. (E) An ag-1 flower with a flowers-within-flower phenotype. (F) An ag-1 clf-47 flower, which was morphologically identical to ag-1 flowers. (G) A sup-1 flower with more stamens than the wild type. (H) An ag-10 clf-47 sup-1 flower, which developed numerous stamens from an indeterminate floral meristem. (I) A knu-1 flower. (J) An ag-10 knu-1 flower with an enlarged gynoecium. (K) An ag-10 clf-47 flower with an enlarged gynoecium. (L) An ag-10 clf-47 knu-1 flower with an internal flower replacing the gynoecium. (M) Siliques from knu-1, ag-10 knu-1, and ag-10 clf-47 plants. The knu-1 silique on the left was a representative silique from young knu-1 plants, while the one on the right was a representative silique from old knu-1 plants. Bars = 1 mm in (A) to (L) and 2.5 mm in (M).

Figure 4.

TFL2/LHP1 Acts in Floral Stem Cell Termination, and WUS Is a PcG Target. (A) A terminal inflorescence composed of several fused flowers in tfl2-2. Note that the gynoecia were thin. (B) A representative inflorescence of ag-10 tfl2-2 plants with floral determinacy defects. The flowers had bulged gynoecia with ectopic floral organs inside (arrow). (C) A representative inflorescence of ag-10 tfl2-2 plants without floral determinacy defects. The gynoecia were thin. (D) Siliques from plants of the indicated genotypes. The ag-10 tfl2-2 plants were from the F2 population of the cross between ag-10 and tfl2-2. Only siliques from the ag-10 tfl2-2 plants with floral determinacy defects are shown. Bars = 1 mm in (A) to (D). (E) A diagram of the WUS genomic region with “+1” being the transcription start site. Gray, black, and white rectangles represent 5′ or 3′ untranslated regions, coding regions, and introns or intergenic regions, respectively. The two red rectangles represent the two CArG boxes. The three regions of TFL2/LHP1 occupancy at WUS as determined by genome-wide profiling of TFL2/LHP1 binding sites are shown in blue (LHP1 DmID; Zhang et al., 2007a). The regions interrogated for AG, H3K27me3, or TFL2/LHP1 enrichment at WUS in this study are shown as black bars. (F) ChIP with anti-H3K27me3 antibodies to determine the levels of H3K27me3 at WUS in wild-type (Ler) and ag-1 inflorescences containing stage 8 and younger flowers. (G) ChIP with anti-HA antibodies in Col (a negative control) and 35S:TFL2-3HA to examine TFL2/LHP1 occupancy at WUS. For (F) and (G), the regions examined are diagramed in (E). eIF4A1 served as a negative control. Error bars represent sd, which were calculated from three technical repeats. Three biological replicates gave similar results.

Figure 5.

AG Binds the WUS locus and Represses WUS Expression Directly. (A) ChIP using anti-AG antibodies to determine AG occupancy at WUS. The null allele ag-1 and the eIF4A1 locus both served as negative controls. AP3, a known direct target of AG (Gómez-Mena et al., 2005), served as a positive control. (B) Real-time RT-PCR to determine WUS transcript levels in 35S:AG-GR ag-1 inflorescences containing stage 8 and younger flowers. Inflorescences were treated with DMSO, DEX, CHX, or CHX plus DEX. Two hours later, the inflorescences were dissected to remove old flowers and harvested for RNA extraction and RT-PCR. Four biological replicates were performed for the DMSO/DEX experiment, and five were performed for the CHX/DEX experiment. Error bars represent sd, which were calculated from these biological repeats. The calculated P values for both experiments were 0.011. (C) Real-time RT-PCR to measure WUS transcript levels in 35S:AG-GR ag-1 clf-47 inflorescences. Chemical treatments and RNA isolation were as in (B).

Figure 6.

Two CArG Boxes within the AG and TFL2/LHP1 Binding Sites at WUS Are Required for the Repression of WUS Expression throughout Flower Development. (A) A diagram of the WUS genomic region as in Figure 4E. The sequences of the region containing the two CArG boxes (capital letters) from Col and Ler as well as the mutated versions are shown. A typical CArG box is CC(A/T)₆GG, but slight variants also serve as functional CArG boxes. (B) A representative inflorescence of WUS1.6:GUS:WUS3′wt transgenic plants showing no GUS staining. (C) A representative inflorescence of WUS1.6:GUS:WUS3′mut transgenic plants showing strong GUS staining in the inflorescence meristem and floral meristems. (D) An inflorescence of a WUS3.2:GUS:WUS3′wt transgenic plant showing GUS staining in the inflorescence meristem and young floral meristems. (E) An inflorescence of a WUS3.2:GUS:WUS3′mut transgenic plant with GUS signals in apparently older flowers than in (D). (F) and (G) Longitudinal sections of an inflorescence (F) or a stage 14 flower (G) of WUS1.6:GUS:WUS3′mut transgenic plants. In (F), the inflorescence meristem (center) is flanked by a stage 1 and a stage 2 floral primordia. GUS signals were present in the inflorescence meristem. In (G), GUS signals were present at the base of the gynoecium. (H) A longitudinal section of a stage 7 WUS3.2:GUS:WUS3′wt flower. This was the latest stage when GUS expression could be detected in this genotype. (I) A longitudinal section of a stage 12 flower from WUS3.2:GUS:WUS3′mut transgenic plants. GUS expression was detected at the base of the gynoecium. Bars = 250 μm in (B), (D), and (E), 400 μm in (C), and 50 μm in (F) to (I).

Figure 7.

AG Recruits PcG to WUS. (A) ChIP using anti-HA antibodies to determine TFL2/LHP1 occupancy at WUS in 35S:TFL2-3HA and 35S:TFL2-3HA ag-1 inflorescences. (B) ChIP using anti-H3K27me3 antibodies in DMSO- or DEX-treated 35S:AG-GR ag-1 inflorescences. At 2 h after treatments, inflorescences were dissected to remove stage 9 and older flowers and used for ChIP. In (A) and (B), real-time PCR reactions were performed with immunoprecipitated and total input DNA. Error bars represent sd, which were calculated from three technical repeats. Three biological replicates gave similar results. The regions interrogated are as diagramed in Figure 4E. eIF4A1 served as a negative control.

Figure 8.

Quantitative Measurements of WUS and AG Expression at Various Stages in Flower Development. Laser capture microdissection was performed to collect cells from the central region of a floral meristem of a defined stage in the wild type ([A] and [B]) and ag-1 ([C] and [D]). Real-time RT-PCR was then performed to examine the levels of WUS ([A] and [C]) and AG ([B] and [D]) transcripts using UBQ5 as the internal control. The levels of expression were shown as relative to those of stage 3, which were set to 1.0. Error bars represent sd, which were calculated from three technical repeats. Two biological replicates gave nearly identical results.

Figure 9.

A Model of the Termination of Floral Stem Cell. AG terminates floral stem cell maintenance by repressing WUS expression (Lenhard et al., 2001; Lohmann et al., 2001). A previous study (Sun et al., 2009) showed that AG represses WUS expression indirectly by activating KNU, which in turn represses WUS expression directly or indirectly. Data presented in this study show that AG also directly represses WUS expression by recruiting PcG to WUS. Genetic studies are consistent with KNU and PcG acting downstream of AG and in parallel to each other in terminating floral stem cell maintenance.