Repression of flowering in Arabidopsis requires activation of FLOWERING LOCUS C expression by the histone variant H2A.Z

Abstract

The histone variant H2A.Z has been implicated in numerous chromatin-mediated processes, including transcriptional activation, euchromatin maintenance, and heterochromatin formation. In yeast and humans, H2A.Z is deposited into chromatin by a conserved protein complex known as SWR1 and SRCAP, respectively. Here, we show that mutations in the Arabidopsis thaliana homologs of two components of this complex, ACTIN-RELATED PROTEIN6 (ARP6) and PHOTOPERIOD-INDEPENDENT EARLY FLOWERING1 (PIE1), produce similar developmental phenotypes and result in the misregulation of a common set of genes. Using H2A.Z-specific antibodies, we demonstrate that ARP6 and PIE1 are required for the deposition of H2A.Z at multiple loci, including the FLOWERING LOCUS C (FLC) gene, a central repressor of the transition to flowering. Loss of H2A.Z from chromatin in arp6 and pie1 mutants results in reduced FLC expression and premature flowering, indicating that this histone variant is required for high-level expression of FLC. In addition to defining a novel mechanism for the regulation of FLC expression, these results support the existence of a SWR1-like complex in Arabidopsis and show that H2A.Z can potentiate transcriptional activation in plants. The finding that H2A.Z remains associated with chromatin throughout mitosis suggests that it may serve an epigenetic memory function by marking active genes and poising silenced genes for reactivation.

Figure 1.

The pie1-5 Mutant Allele Is Null. (A) Diagram of the PIE1 gene. The transcription start site is shown as a right-facing arrowhead, and the termination site is indicated by an X. Exons are depicted as gray boxes. The location of the T-DNA insertion in pie1-5 is shown as a triangle above exon 15. (B) Relative quantity of the PIE1 transcript in wild-type and pie1-5 plants as detected by real-time RT-PCR. The transcript is essentially undetectable in the pie1-5 mutant, indicating that the allele is null. Error bars represent se (n = 3).

Figure 2.

arp6 and pie1 Mutants Exhibit Similar Phenotypes. (A) Twenty-day-old wild-type, arp6-1, and pie1-5 plants grown under long-day conditions. (B) Real-time RT-PCR data showing relative expression of the indicated genes in wild-type, arp6-1, and pie1-5 plants. Average relative quantities ± se (n = 3) are shown for each sample. (C) Protein gel blots of even-numbered gel filtration fractions from a Sephacryl S-300 column run with extract from wild-type (top panel) or pie1-5 (bottom panel) plants. Blots were probed with an ARP6 monoclonal antibody. The input lanes in each blot were loaded with ∼25 μg of the unfractionated protein extract from each genotype. Asterisks indicate the ARP6 peak fractions, and calibrated molecular masses are given below the blots.

Figure 3.

H2A Phylogeny and H2A.Z-Specific Antibodies. (A) Neighbor-joining protein sequence phylogeny showing H2A proteins from Arabidopsis (At), rice (Os), Drosophila (Dm), human (Hs), and Saccharomyces (Sc). Bootstrap values are shown on the branch points of the tree. The H2A.Z clade is indicated by a vertical bar at right, and asterisks indicate proteins used in (B) and (C). (B) Coomassie blue–stained SDS-PAGE gel showing purified recombinant Arabidopsis H2A histones. The H2A.Z variants are indicated by a horizontal bar above the protein names. Molecular masses are shown at right. (C) Protein gel blot of a gel loaded as in (B), probed with the H2A.Z-specific antibody. Molecular masses are shown at right. The polyclonal antibody (pAb) recognized HTA9 and HTA11 and was predicted to react with HTA8, because it is nearly identical to HTA11 in the N-terminal region. The HTA4 protein is highly divergent from the other H2A.Zs at the N terminus and is not expected to be recognized by the antibody. (D) Protein gel blot of immunoprecipitates probed with an ARP6 monoclonal antibody (mAb). The input sample was 5% of the total protein used in each immunoprecipitation. The antibody used for immunoprecipitation is indicated above the blot: either the H2A.Z antibody or preimmune serum (PI) from the same rabbit.

Figure 4.

H2A.Z Localizes to Euchromatic Regions but Not to Heterochromatic Chromocenters. (A) Isolated leaf cell nucleus probed with the H2A.Z antibody. (B) The 4′,6-diamidino-2-phenylindole (DAPI) channel image of the nucleus shown in (A). Chromocenters appear as densely stained spots throughout the nucleus. (C) Merge of images shown in (A) and (B). (D) Anaphase-stage root cell probed with the H2A.Z antibody. (E) DAPI channel image of the cell shown in (D). (F) Merge of images shown in (D) and (E).

Figure 5.

ARP6 and PIE1 Are Required for the Deposition of H2A.Z at FLC, MAF4, and MAF5. (A) Enrichment of H2A.Z on the FLC gene in the wild type and mutants as measured by ChIP with the H2A.Z antibody. The graph shows average fold enrichment ± se as measured by real-time PCR. (B) Enrichment of H2A.Z on the MAF4 and MAF5 genes in the wild type and mutants as measured by ChIP and real-time PCR. (C) Enrichment of H2B on the FLC gene in the wild type and arp6-1. (D) Enrichment of H2B on the MAF4 and MAF5 genes in the wild type and arp6-1. For each ChIP experiment ([A] to [D]), n = 3. (E) Diagram of the FLC gene with exons indicated as gray boxes. The transcription start site is shown as an arrowhead, and the termination site is shown as an X. PCR primer sets are shown as black boxes below the diagram. Primer set numbers correspond to the numbers on the x axis of the graphs in (A) and (C). (F) Diagram of the MAF4 and MAF5 genes and locations of PCR primer sets, depicted as in (E). Primer set numbers correspond to the numbers on the x axis of the graphs in (B) and (D).

Figure 6.

Relationship between FLC Expression Levels and H2A.Z Abundance on the Gene. (A) Real-time RT-PCR results showing the relative FLC level in each genotype or tissue. FRI indicates a Columbia line carrying the strong FRI allele from the Sf-2 ecotype (Lee and Amasino, 1995), WT indicates Columbia wild type, and CL indicates cauline leaf. The graph shows average relative quantities ± se (n = 3). (B) Enrichment of H2A.Z on the FLC gene in the indicated samples as measured by ChIP and real-time PCR. The graph shows average fold enrichment ± se. (C) ChIP was performed on the indicated samples using either an H2A.Z antibody or an H2B antibody, and data are reported as H2A.Z enrichment/H2B enrichment to correct the H2A.Z levels for total nucleosome occupancy. The graph shows average fold enrichment ± se. For each ChIP experiment ([B] and [C]), n = 3. (D) Diagram of the FLC gene with exons indicated as gray boxes. The transcription start site is shown as an arrowhead, and the termination site is shown as an X. PCR primer sets are shown as black boxes below the diagram. Primer set numbers correspond to the numbers on the x axis of the graphs in (B) and (C).