Resetting the Yeast Epigenome with Human Nucleosomes

Abstract

Humans and yeast are separated by a billion years of evolution, yet their conserved histones retain central roles in gene regulation. Here, we "reset" yeast to use core human nucleosomes in lieu of their own (a rare event taking 20 days), which initially only worked with variant H3.1. The cells adapt by acquiring suppressor mutations in cell-division genes or by acquiring certain aneuploid states. Converting five histone residues to their yeast counterparts restored robust growth. We reveal that humanized nucleosomes are positioned according to endogenous yeast DNA sequence and chromatin-remodeling network, as judged by a yeast-like nucleosome repeat length. However, human nucleosomes have higher DNA occupancy, globally reduce RNA content, and slow adaptation to new conditions by delaying chromatin remodeling. These humanized yeasts (including H3.3) pose fundamental new questions about how chromatin is linked to many cell processes and provide a platform to study histone variants via yeast epigenome reprogramming.

Keywords: chromatin; genomics; histone; histones; humanized; synthetic biology; systems biology.

Figure 1. Saccharomyces cerevisiae can subsist on completely human core nucleosomes

(A) Human and budding yeast histone differences. Red bars indicate residue positions that differ between the two species. Numbers refer to the yeast histones. Gray colored regions show the globular histone domains, and the white regions show the N- and C-terminal tails. (B) Nucleosome deposition schematic and variants from yeast and humans. (C) Dual-histone plasmid shuffle strategy (see Figure S1). (D) The “humanization frequency” by which each human histone gene in the context of the relevant histone pairs or all four histone genes complement deletions of the respective yeast counterpart was assessed as in (C). Spots show yeast 10-fold serial dilutions. The densest spot contains 10⁵ yeast. (E) Only 8 yeast colonies with completely human nucleosomes (referred to as “yHs”) arise after 20 or more days on plates. (F) PCRtag analysis of humanized yeast, and confirmed by sequencing extracted plasmids.

Figure 2. Acquisition of bypass mutations in cell-cycle genes promotes growth with human nucleosomes

(A) Seven yHs-strains were evolved for 5 cycles in liquid medium (SC–Trp). (B) Growth rates and doubling times (d) in SC–Trp. Complete doubling times are listed in Table S2. (C) Types of DNA mutations identified by whole-genome sequencing (WGS). (D) Number of times each listed chromosome was scored as aneuploid. (E) 22 unique mutations identified by WGS (Tables S1 and S2) were constructed into a network (p=5×10⁻⁵) using the String algorithm. Colored nodes are in similar processes. Black nodes are top 4 interacting genes inferred from the network, but not arising as suppressors. (F) Transcription factors common amongst the 22 suppressors. Numbers in parentheses = number of genes identified by suppressor mutations with binding sites in promoters.

Figure 3. Specific residues in the C-termini of histones H3 and H2A limit growth rates

(A) Maps of swap-back residues that enhance human histone utilization identified in Figures S3 and S4. Two residues in the C-terminus of human histone H3 (hH3), and three swap-back residues each in the N-terminus or C-terminus of human histone H2A (hH2A) improved the complementation frequency and growth rate in conjunction with their respective human histone counterpart (i.e., hH4 and hH2B respectively). (B) Systematic combinations of swap-back residues in hH3 and hH2A, along with fully human hH4 and hH2B show that eight swap-back residues promote the highest rates of complementation. (C) Colony growth rate analyses shows that the five-residue “swap-back” strain (yDT97) grows as well as the eight-residue “swap-back” strain (yDT98). Both swap-back strains grow at rates closer to isogenic-WT yeast (yDT67), and better than the fastest growing completely humanized isolate (yHs5-evo). All further swap-back strain experiments refer to yDT97.

Figure 4. Human nucleosome organization in yeast

(A) MNase digestions reveal that human nucleosomes produce the same nucleosome repeat length as yeast nucleosomes, compared to the longer length of human nucleosomes in HeLa cells. Red arrows indicate position of the tri-nucleosome. The “bp” indicates base-pair size of the DNA ladder (“L”). (B) Table of high (2 units/ml) MNase-seq nucleosome dynamics between humanized to WT yeast, and WT experiment 1 to WT experiment 2 (“noise”). Occupancy and fuzziness changes use a strict False Discovery Rate cut-off of 0.05 (p < 10⁻⁸⁵) and additional parameters in STAR Methods. (C) High MNase-seq read counts at centromeric regions, plotted for chromosomes that were normal or aneuploid in Figure 2D. RCPM refers to read counts per million mapped reads. (D) High MNase-seq read counts for all 275 tRNA genes comparing humanized vs. WT strains showing depletion of either RNAP3 or nucleosomes.

Figure 5. Human nucleosomes are more repressive

(A) Pre-evolved yHs strains have reduced levels of bulk total RNA (6–8 fold), as well as the evolved and swap-back strains (2–3 fold). Data show mean ± SEM of 3 biological replicates. (B) Acid-extracted histones from strains analyzed for equal loading by Coomassie staining, and then immunoblotted using different H3 and H4 antibodies. (C) Volcano plot of differentially expressed genes shared amongst all 7 pre-evolved yHs strains compared to isogenic-WT yeast. 1046 genes had a q-value <10⁻⁴, and group into the GO terms described. The q-value is the False Discovery Rate adjusted p-value. (D) Heatmaps and average profiles of high concentration MNase-seq reads aligned around the transcription start sites (TSS) ±500 bp of the top and bottom 1500 genes by expression. RCPM and color key refers to read counts per million mapped reads. NDR refers to the nucleosome depleted region. (E) Repressed genes with highly occupied −1 nucleosomes from MNase-seq.

Figure 6. Humanized yeast have trouble adapting to new conditions

(A) Violin plots showing that humanized yeast cells are larger and have dysregulated cell size based on phase-contrast microscopy measurements. (B) Cell-cycle (CLB1) and cell-size (WHI4) regulating genes each have highly occupied −1 and −2 nucleosomes. (C) Humanized yeast have a prolonged S-phase and/or arrest in G1. Cell-cycle was analyzed by sytox green straining of DNA content and measured by flow cytometry. Each plot shows 10,000 cells. (D) Humanized yeast have delayed remodeling at the GAL1 promoter. Time-course was analyzed by galactose induction of eGFP using flow cytometry. (E) MNase-seq map of PHO5 promoter, and time-course nucleosome scanning assay using WT, hH3.1-core, and hH3.3-core nucleosome yeasts upon phosphate starvation at different time points. Data points show qPCR amplicon midpoints, and mean ± SD of 2 biological replicates.

Figure 7. “Re-humanization” of suppressor mutants starting with human or yeast chromatin, and generation of an H3.3 humanized strain

(A) Suppressor strains were “re-humanized” (see (C)), by generating a mixed chromatin environment (yeast + human), either starting from human chromatin (red dots) or native yeast chromatin (black dots). Genotypes are listed at top, and then assayed by the dual-plasmid shuffle experiment. Each dot represents the log10 mean of 3 or more experiments ± SEM. Suppressor strain with red asterisk was unable to lose human histone plasmid. (B) Humanization frequency of hH3.3-core nucleosome, log10 mean of 8 experiments ± SEM. (C) Diagram of suppressor strain “re-humanization” frequency experiments. The dotted line inset shows the possible chromatin states that may apply when yeast and human nucleosomes coexist and partition to new cells.