A key role for Chd1 in histone H3 dynamics at the 3' ends of long genes in yeast

Abstract

Chd proteins are ATP-dependent chromatin remodeling enzymes implicated in biological functions from transcriptional elongation to control of pluripotency. Previous studies of the Chd1 subclass of these proteins have implicated them in diverse roles in gene expression including functions during initiation, elongation, and termination. Furthermore, some evidence has suggested a role for Chd1 in replication-independent histone exchange or assembly. Here, we examine roles of Chd1 in replication-independent dynamics of histone H3 in both Drosophila and yeast. We find evidence of a role for Chd1 in H3 dynamics in both organisms. Using genome-wide ChIP-on-chip analysis, we find that Chd1 influences histone turnover at the 5' and 3' ends of genes, accelerating H3 replacement at the 5' ends of genes while protecting the 3' ends of genes from excessive H3 turnover. Although consistent with a direct role for Chd1 in exchange, these results may indicate that Chd1 stabilizes nucleosomes perturbed by transcription. Curiously, we observe a strong effect of gene length on Chd1's effects on H3 turnover. Finally, we show that Chd1 also affects histone modification patterns over genes, likely as a consequence of its effects on histone replacement. Taken together, our results emphasize a role for Chd1 in histone replacement in both budding yeast and Drosophila melanogaster, and surprisingly they show that the major effects of Chd1 on turnover occur at the 3' ends of genes.

Figure 1. Chd1 affects H3.3_core-GFP localization on Chd1 in Drosophila.

(A) Representative sections from confocal imaging of H3.3_core-GFP in nuclei from salivary glands of wild type larvae, (B) chd1⁵ heterozygotes and (C) chd1⁵ homozygotes. The GFP signal is pseudo green. In all cases, H3.3_core-GFP was expressed from P[UHS-H3.3_core-GFP] and driven by P{GawB} AB1-Gal4. (D) Quantitation of banding patterns observed in nuclei from flies with the indicated genotypes. A total of 44 wild type, 144 heterozygote, and 162 homozygous null nuclei were scored, all blind to genotype.

Figure 2. The H3 N-terminal tail functions redundantly with Chd1 and an H3.3-like surface of histone H3 in budding yeast.

The indicated histone H3 plasmids (which also carried histone H4) were transformed into wild type CHD1 or chd1 null strains that lack both chromosomal copies of the histone H3/H4 genes and contained a URA3 H3/H4 plasmid. Cultures were adjusted to 1×10⁷ cells per ml and five-fold serial dilutions were spotted directly onto 5FOA media, selecting for cells that had lost the URA3 H3/H4 plasmid, and incubated for 2 days at 30°C.

Figure 3. Chd1 plays a key role in H3 replacement dynamics at gene ends.

(A) H3 replacement was measured in G1-arrested cells by induction of Flag-H3 for 60 minutes, followed by ChIP enrichment of both Flag-containing and total H3-associated DNA and subsequent competitive hybridization to tiling microarrays. H3 turnover is represented as log2 of Flag-H3 ChIP enrichment over total-H3 ChIP (y axis). Here, data for all yeast genes is shown in a “metagene” view, with 10 bins of 50 bp each, upstream of the +1 nucleosome, followed by 20 bins representing 5% increments along gene coding regions, scaling for gene length. Three independent replicate experiments are shown along with the averaged profile, as indicated. (B) Average Flag-H3/total-H3 for various classes of genomic element, defined as in , . Briefly, 5′ and 3′ CDS refer to probes within the first and last 500 bp of coding regions, with mid-CDS encompassing any remaining probes. TSS indicated probes up to 500 bp upstream of the ATG, and Promoter includes all remaining upstream probes. ARS includes all probes within 200 bp of an ARS. Null indicates all remaining probes, predominantly those that fall between convergently-transcribed genes.

Figure 4. Length dependence of Chd1 effects on H3 replacement.

H3 replacement was averaged for the 500 bp at the 5′ ends of genes (A), or the 3′ ends of genes (B). Genes were ordered by length, and an 80 gene window average is shown for wild type and chd1Δ turnover data as indicated. Bottom panel plots gene lengths, and locations for 1, 2, and 3 kb are indicated below panel (B).

Figure 5. Chd1 effects on H3 methylation patterns.

H3K4me3 and H3K36me3 were mapped genome-wide by ChIP-chip on tiling microarrays. Metagene analysis is shown for wild type and chd1Δ strains, as indicated.