1. Chromosomes and Gene Expression
  2. Genetics and Genomics
Download icon

The ACF chromatin-remodeling complex is essential for Polycomb repression

  1. Elizabeth T Wiles  Is a corresponding author
  2. Colleen C Mumford
  3. Kevin J McNaught
  4. Hideki Tanizawa
  5. Eric U Selker  Is a corresponding author
  1. Institute of Molecular Biology, University of Oregon, United States
Research Article
  • Cited 0
  • Views 991
  • Annotations
Cite this article as: eLife 2022;11:e77595 doi: 10.7554/eLife.77595

Abstract

Establishing and maintaining appropriate gene repression is critical for the health and development of multicellular organisms. Histone H3 lysine 27 (H3K27) methylation is a chromatin modification associated with repressed facultative heterochromatin, but the mechanism of this repression remains unclear. We used a forward genetic approach to identify genes involved in transcriptional silencing of H3K27-methylated chromatin in the filamentous fungus Neurospora crassa. We found that the N. crassa homologs of ISWI (NCU03875) and ACF1 (NCU00164) are required for repression of a subset of H3K27-methylated genes and that they form an ACF chromatin-remodeling complex. This ACF complex interacts with chromatin throughout the genome, yet association with facultative heterochromatin is specifically promoted by the H3K27 methyltransferase, SET-7. H3K27-methylated genes that are upregulated when iswi or acf1 are deleted show a downstream shift of the +1 nucleosome, suggesting that proper nucleosome positioning is critical for repression of facultative heterochromatin. Our findings support a direct role of the ACF complex in Polycomb repression.

Editor's evaluation

This manuscript provides strong evidence that ACF directly functions to promote Polycomb-dependent repression through chromatin remodeling, which has not been demonstrated. In addition, PRC2/H3K27me-dependent ACF targeting is novel. Finally, the authors' model that facultative chromatin can be classified into several groups based on their dependence on SET7, ASH1, and ACF (Figure 6) is potentially important for guiding future research directions of the field.

https://doi.org/10.7554/eLife.77595.sa0

eLife digest

All the cells in an organism contain the exact same DNA, yet each type of cell performs a different role. They achieve this by turning specific genes on or off. To do this, cells wind their genetic code into structures called nucleosomes, which work a bit like spools of thread. Chemical modifications on these nucleosomes can determine whether a cell will use the genes spooled around it or not.

In many organisms, cells can turn genes off using a modification called H3K27 methylation. This mark attracts a protein complex called PRC1 that packs the genes away, making them inaccessible to the proteins that would activate them. But the filamentous fungus Neurospora crassa does not produce PRC1. This suggests that this organism must keep genes with the H3K27 mark switched off in a different way. One possibility is that H3K27 methylation somehow leads to changes in the position of nucleosomes on the genome, since having nucleosomes near the beginning of gene sequences can stop the cell from reading the code.

One protein complex responsible for positioning nucleosomes is known as the ATP-utilizing chromatin assembly and remodeling factor (ACF) complex, but it remained unknown whether it interacted with H3K27 methylation marks. To investigate further, Wiles et al. generated strains of Neurospora crassa that did not synthesize ACF and discovered that many of their genes, including ones marked with H3K27, were turned on. This was probably because the nucleosomes had shifted out of position, allowing the proteins responsible for activating the genes to gain access to the start of the genes’ sequences.

Turning genes on and off at the right time is crucial for development, cell survival, and is key in tissues and organs working properly. Understanding the role of ACF adds to what we know about this complex process, which is involved in many diseases, including cancer.

Introduction

Polycomb repressive complex 2 (PRC2) methylates lysine 27 of histone H3 (H3K27), marking facultative heterochromatin (Müller et al., 2002; Margueron and Reinberg, 2011). Facultative heterochromatin contains regions of the genome that must remain transcriptionally plastic in order to respond to developmental or environmental cues (Wiles and Selker, 2017). Although H3K27 methylation has been established as a repressive chromatin modification, the mechanisms of repression are not fully understood (Margueron and Reinberg, 2011; Ridenour et al., 2020). One model for repression involves PRC1 binding to the H3K27 methyl-mark to facilitate chromatin compaction by self-association (Grau et al., 2011; Cheutin and Cavalli, 2018; Boyle et al., 2020). However, PRC1 is not present in all eukaryotes that bear H3K27 methylation-associated silencing, such as the filamentous fungus Neurospora crassa (Jamieson et al., 2013; Schuettengruber et al., 2017; Wiles et al., 2020), suggesting the existence of additional mechanisms of H3K27 methylation-associated repression.

In addition to histone modifications, nucleosome positioning may be critical for facultative heterochromatin function. The nucleosome, which is the fundamental unit of chromatin, consists of approximately 147 base pairs of DNA wrapped around an octamer of histone proteins (Kornberg, 1974; Luger et al., 1997). Nucleosomes can be precisely positioned on DNA by ATP-dependent chromatin-remodeling proteins to produce a chromatin landscape that modulates accessibility for DNA transactions, such as transcription (Lai and Pugh, 2017). In particular, the precise positioning of the +1 nucleosome, the first nucleosome downstream of the transcription start site (TSS), is thought to be an important determinant of gene expression (Rhee and Pugh, 2012; Nocetti and Whitehouse, 2016). This dynamic nucleosome can occlude binding elements for transcriptional regulatory sites such as the TATA box (Kubik et al., 2018) and can serve as a barrier to RNA polymerase II (Weber et al., 2014).

In Drosophila melanogaster, the ATP-utilizing chromatin assembly and remodeling factor (ACF) complex (Ito et al., 1997) has been indirectly linked to the repression of Polycomb targets (Fyodorov et al., 2004; Scacchetti et al., 2018). The ACF complex is composed of the ATPase, Imitation Switch (ISWI), and the accessory subunit ACF1 (Ito et al., 1999). ACF is thought to act as a global nucleosome spacer and to contribute to repression genome-wide (Baldi et al., 2018; Scacchetti et al., 2018). Mutations in Acf1 act as enhancers of Polycomb mutations and disrupt nucleosome spacing in facultative heterochromatin (Fyodorov et al., 2004; Scacchetti et al., 2018).

In order to improve our understanding of the control and function of facultative heterochromatin, we used forward genetics to identify factors required for silencing H3K27-methylated genes in N. crassa. As described here, this identified iswi (also known as crf4-1) and acf1 (also known as crf4-2; Itc1 in S. cerevisiae). We show that these proteins interact to form an ACF complex in N. crassa, that ACF is necessary for repression of a subset of H3K27-methylated genes, and that derepression is not simply due to loss of H3K27 methylation. ACF interacts with chromatin targets throughout the genome, yet specific interactions with H3K27-methylated regions are partly dependent on SET-7, the H3K27 methyltransferase. Finally, we show that when members of ACF are absent, H3K27-methylated genes that become upregulated display a specific downstream shift of the +1 nucleosome. Our findings support a model in which ACF remodels the chromatin landscape at H3K27-methylated regions of the genome to contribute to Polycomb silencing.

Results

Forward genetic selection for genes required for Polycomb silencing identifies iswi and acf1

We previously designed and employed a forward genetic selection to identify novel genes required for silencing of H3K27-methylated genes (McNaught et al., 2020; Wiles et al., 2020). Briefly, the open reading frames of two genes (NCU05173 and NCU07152) that require the H3K27 methyltransferase (SET-7) for repression were replaced with the hygromycin (hph) and nourseothricin (nat-1) resistance genes, respectively (Figure 1A). The resulting strain, which was sensitive to both antibiotics, was subjected to UV mutagenesis and grown on hygromycin- and nourseothricin-containing medium to select for drug-resistant colonies. As a step to identify mutations causing drug resistance, these strains, which were in an Oak Ridge genetic background, were crossed to the polymorphic Mauriceville strain (Metzenberg et al., 1985). Spores from these crosses were germinated on hygromycin- and/or nourseothricin-containing media to select for the mutant segregants, and genomic DNA from the progeny was pooled for whole-genome sequencing. Critical mutations were mapped by examining the ratio of Oak Ridge to Mauriceville single-nucleotide polymorphisms (SNPs) genome-wide (Pomraning et al., 2011). Genetic variants were identified within the mapped regions using published tools (Danecek et al., 2011; Garrison and Marth, 2012).

Figure 1 with 1 supplement see all
Forward genetics identifies ISWI complex members required for repression of H3K27-methylated genes.

(A) Selection scheme with reporter genes inserted at H3K27 methylation-marked loci to select for genes required for silencing. (B) Schematic of protein domains in ISWI and ACF1 with the changes identified in our selection (L430P and D161fs, respectively; marked with asterisks). The conserved nature of the changed residue in ISWI is highlighted for the designated species. (C) Serial dilution spot-test silencing assay for the indicated strains, which all contain PNCU07152::nat-1 on media with or without nourseothricin (NTC). (D), Serial dilution spot-test silencing assay for the indicated strains, which contain PNCU07152::nat-1 and PNCU5173::hph, on media with or without nourseothricin (NTC) or hygromycin (HYG). For complementation tests, wild-type copies of each gene were inserted at the his-3 locus (indicated at left as +iswiWTor +acf1WT). All spot tests were imaged after 48 hr at 32°C and performed at least twice. The number of cells spotted is indicated beneath the images.

SNP mapping for one mutant identified a region on linkage group VI that contained essentially 100% Oak Ridge SNPs, indicating the likely position of the critical mutation (Figure 1—figure supplement 1A). Within this region, we found a point mutation (CTT → CCT) predicted to cause a leucine to proline substitution at a conserved position in iswi (NCU03875) (L430P; Figure 1B). This same approach was used on a second mutant to map and identify a two base pair insertion (GTC → TTGTC) on linkage group III (Figure 1—figure supplement 1B) that leads to a frameshift (D161fs) in acf1 (NCU00164) (Figure 1B).

To test if deletion of these two identified genes would also cause derepression, we created strains with the NCU07152::nat-1 replacement and either ∆iswi or ∆acf1 alleles. Indeed, deletion of either iswi or acf1 resulted in nourseothricin-resistance, equivalent to the original mutants identified in our selection (Figure 1C). In addition, we showed that introduction of an ectopic, wild-type copy of iswi or acf1 into the corresponding original mutant strain largely restored silencing of both H3K27 methylation mutant selection genes (Figure 1D). We noticed that disruption of iswi or acf1 resulted in an early conidiation phenotype (production of asexual spores) that appeared as more dense growth in the spot tests (Figure 1C and D), but this was not accompanied by an increased linear growth rate. In fact, the ∆iswi strain showed a decreased growth rate relative to wild type; the ∆acf1 strain grew comparably to wild type (Figure 1—figure supplement 1C). Taken together, these data confirm that iswi and acf1 are required to maintain the drug sensitivity of strains containing the H3K27 methylation mutant selection genes and are thus good candidates for genes involved in repression of H3K27-methylated chromatin.

ISWI and ACF1 form a complex in N. crassa

Evidence from several organisms, most notably budding yeast and Drosophila, has implicated ISWI as the catalytic subunit of several chromatin-remodeling complexes (Petty and Pillus, 2013). To look for possible ISWI-containing protein complexes in N. crassa, we affinity-purified overexpressed 3xFLAG-ISWI from N. crassa cellular extracts. Immunopurified samples were digested down to peptides and analyzed by mass spectrometry (MS) to identify potential interacting proteins. We focused on proteins whose counts comprised greater than 0.4% of the total spectrum counts. ISWI co-purified with ACF1 as well as with CRF4-3 (NCU02684), a homolog of Ioc4 and member of the Saccharomyces cerevisiae Isw1b complex (Vary et al., 2003). Top hits from the MS results also included NCU00412 and NCU09388, proteins not known from S. cerevisiae (using NCBI BLASTP Altschul et al., 1990; Figure 2, Figure 2—figure supplement 1). Nevertheless, we considered that these could be members of ISWI complexes based on the high number of unique peptides detected and the presence of a WHIM domain in NCU00412 and a PHD domain in NCU09388—domains that are present in S. cerevisiae Isw1 complex members, Ioc3 and Ioc2, respectively (Vary et al., 2003). NCU00412 and NCU09388 were also identified in a recent independent analysis of ISWI-interacting proteins in N. crassa and named ISWI accessory factors 1 and 2 (IAF-1 and IAF-2), respectively (Kamei et al., 2021). CRF4-3 was not previously identified as an ISWI-interacting partner, but for consistency, we will adopt the new nomenclature and refer to this protein as ISWI accessory factor 3 (IAF-3).

Figure 2 with 2 supplements see all
ISWI and ACF1 interact in Neurospora crassa and are required for repression of a subset of SET-7-repressed genes.

(A) Schematic representation of ISWI-interactions found by immunoprecipitation followed by mass spectrometry. Proteins (ISWI/NCU03875, ACF1/NCU00164, IAF-3/NCU02684, IAF-1/NCU00412, IAF-2/NCU09388, HFP-1/NCU03073, and HFP-2/NCU06623) are depicted to scale. Arrows are drawn from the protein used as the ‘bait’ to the protein partner identified, and unique peptide counts are indicated. Dotted arrows indicate the peptide count was below 0.4% of the total spectrum threshold. Proteins in gray (HFP-1 and HFP-2) were identified as interacting partners but were not used as ‘bait.’ (B) Serial dilution spot-test silencing assay for the indicated strains on media with or without nourseothricin (NTC). All strains have PNCU07152::nat-1. The number of cells spotted is indicated beneath the images, which were generated after incubation for 48 hr at 32°C. Spot test assays were repeated at least twice. (C–G) Expression level (FPKM) for each gene in the indicated mutant strain plotted against the expression level in wild type. Two biological replicates were used for each mutant. Two biological replicates were perfomed twice for wild type. Differentially expressed (DE) genes were defined using a significance cutoff of log2fold change>2 for upregulated genes and log2fold change<–2 for downregulated genes with a p value <0.05. Gray dots indicate genes that are not considered DE. Upper left corner shows the total number of significantly up- and downregulated genes with the number of H3K27-methylated genes in parentheses. Significance for enrichment of H3K27-methylated genes in each DE gene set was calculated by Fisher’s exact test (FPKM - fragments per kilobase per million reads). (H) Venn diagram showing overlap between H3K27-methylated genes that are upregulated (log2fold change>2; p value <0.05) in ∆iswi and ∆acf1 strains. Significant overlap (p<8.036E−54) determined by hypergeometric probability test.

To confirm these interactions and to gain information on the possible formation of ISWI-containing subcomplexes, we engineered a C-terminal HA tag at the endogenous locus of each of the four most prominent putative ISWI-interacting partners: ACF1, IAF-3, IAF-1, and IAF-2. These proteins were purified by immunoprecipitation and subjected to MS to identify interacting partners. Interactions between ISWI and all four proteins were confirmed, with each HA-tagged protein pull-down yielding high unique peptide counts for ISWI. Additional interactions, with lower unique peptide counts, and typically lack of reciprocal pull-downs, were also found (Figure 2A and Figure 2—figure supplement 1A). These data suggest that ISWI forms multiple distinct protein complexes and, importantly, that ISWI and ACF1, two proteins identified in our selection for factors involved in the repression of H3K27-methylated genes, interact. The ACF1-HA pull-down identified two histone fold proteins, NCU03073 (HFP-1) and NCU06623 (HFP-2) (Borkovich et al., 2004; Kamei et al., 2021), as interacting partners (Figure 2A and Figure 2—figure supplement 1A). These proteins are notable because histone fold proteins are found in the CHRAC complex (DPB4 and DLS1 in S. cerevisiae and CHRAC14/16 in D. melanogaster) along with ISWI and ACF1 (Varga-Weisz et al., 1997; Corona et al., 2000; Iida and Araki, 2004).

To investigate if any of the identified ISWI-interacting proteins, beyond ACF1, are involved in H3K27-methylated gene silencing, we first examined whether they are required for silencing the NCU07152::nat-1 selection marker. As previously shown, deletion of iswi or acf1 results in robust growth on nourseothricin, indicating strong derepression of the nat-1 gene. Deletion of iaf-3 or hfp-1 also derepressed the nat-1 marker. Strains with deletion of iaf-1 and iaf-2 showed more modest growth on nourseothricin (Figure 2B). These data show that ISWI, ACF1, and other ISWI-interacting proteins contribute to the silencing of the NCU07152::nat-1 selection marker.

iswi and acf1 are required for repression of a subset of H3K27-methylated genes

We performed mRNA-seq on ∆iswi and ∆acf1 strains to determine if the loss of these genes affects transcription beyond the NCU07152::nat-1 selection marker and, if so, to determine if these effects were specific to H3K27-methylated domains, or were more general. We found that while the majority of gene expression changes observed upon loss of ISWI or ACF1 occurred outside of H3K27-methylated domains (Figure 2—figure supplement 2A, B), genes marked by H3K27 methylation were significantly enriched in the upregulated gene sets for ∆iswi and ∆acf1 strains (Figure 2C and D). To determine the extent to which the other three ISWI-interacting partners contribute to silencing in H3K27-methylated regions, we performed mRNA-seq on strains with deletions of iaf-3, iaf-1, or iaf-2. We found that H3K27-methylated genes were modestly enriched in the ∆iaf-3 and ∆iaf-2 gene sets (Figure 2E and F) but not enriched in the ∆iaf-1 gene set (Figure 2G). Nearly all (92%) of the H3K27-methylated genes that were upregulated in ∆acf1 were also upregulated in ∆iswi, showing significant (p<8.036E−54) overlap between these two gene sets (Figure 2H). Only 30% of these genes were part of the ∆set-7 upregulated gene set (Figure 2—figure supplement 2C), consistent with the notion that the repression of H3K27-methylated genes is not solely a result of PRC2 activity. This demonstrates that ISWI and ACF1 are not simply involved in the repression of the two H3K27-methylated genes that we used in our initial selection (NCU05173 and NCU07152) but are also necessary for the repression of a large overlapping set of H3K27-methylated genes.

iswi is required for wild-type H3K27 methylation and H3K36 trimethylation

We know that loss of H3K27 methylation (Jamieson et al., 2013) or loss of H3K36 methylation (Bicocca et al., 2018) is associated with derepression of genes in facultative heterochromatin. To investigate if upregulation of genes in facultative heterochromatin in ∆iswi and ∆acf1 strains is due to loss of H3K27- or H3K36-methylation in these regions, we performed ChIP-seq for H3K27me2/3, H3K36me2, and H3K36me3. We compared the level of each of these histone modifications in ∆iswi and ∆acf1 strains to that of wild type. We found that changes in H3K36me2 in each of these mutants were negligible (Figure 3—figure supplement 1A, B). We saw minor loss of both H3K27me2/3 and H3K36me3 in ∆acf1, and more changes in these histone marks in ∆iswi strains (Figure 3A–D). H3K27me2/3 ChIP-seq of strains with genes for other ISWI-interacting proteins deleted (∆iaf-3, ∆iaf-1, or ∆iaf-2) showed only minor changes in H3K27me2/3 (Figure 3—figure supplement 2A-C).

Figure 3 with 2 supplements see all
iswi and acf1 are required for wild-type H3K27me2/3 and H3K36me3 but loss of these methyl marks is not required for transcriptional upregulation.

(A, B) Scatter plots show the correlation of H3K27me2/3 at genes in wild type and ∆iswi or ∆acf1 based on biological replicates of ChIP-seq data. Green points (n=260 in ∆iswi and n=0 in ∆acf1) represent genes with increased H3K27me2/3 levels (at least twofold over wild type) and red points (n=341 in ∆iswi and n=193 in ∆acf1) represent genes with decreased H3K27me2/3 levels (at least twofold relative to wild type) in the indicated mutant. (C, D) Scatter plots show the correlation of H3K36me3 at genes in wild type and ∆iswi or ∆acf1 based on biological replicates of ChIP-seq data. Green points (n=1 in ∆iswi and n=0 in ∆acf1) represent genes with increased H3K36me3 levels (at least twofold over wild type) and red points (n=444 in ∆iswi and n=317 in ∆acf1) represent genes with decreased H3K36me3 levels (at least twofold relative to wild type) in the indicated mutant. (E, F) Scatter plots show the correlation between H3K27me2/3 and gene expression at H3K27-methylated genes (n=836) in the indicated mutants. Pearson correlation coefficient is reported. Red box indicates genes (n=92 in ∆iswi and n=66 in ∆acf1) that are significantly upregulated (log2 fold change>2) but show no significant loss of H3K27me2/3 (log2 fold change>–1). (G, H) Scatter plots show the correlation between H3K36me3 and gene expression at H3K27-methylated genes (n=836) in the indicated mutants. Pearson correlation coefficient is reported. (I) ChIP-seq tracks showing average level of H3K27me2/3 or H3K36me3 merged from two biological replicates for the indicated strains on LG III. Y-axis is 0–1000 RPKM for H3K27me2/3 tracks and 0–100 average read counts for H3K36me3 tracks. (J) Same as in (I), but for LG IV. (K) Enlarged ChIP-seq tracks showing the underlined region on LG IV from (J). Gene expression changes in ∆iswi are shown. (L) ChIP-qPCR data for H3K27me2/3 at the two genes used for the initial mutant selection (NCU05173 and NCU07152) in the indicated strains. Filled bars represent the mean of technical triplicates and error bars show standard deviation (** for p<0.01, * for p<0.05, and ns for not significant; all relative to wild type by unpaired t-test). Data are from one representative experiment that was performed three times.

Figure 3—source data 1

H3K27me2/3 ChIP-seq comparisons (∆iswi and ∆acf1).

https://cdn.elifesciences.org/articles/77595/elife-77595-fig3-data1-v3.xlsx
Figure 3—source data 2

H3K36me3 ChIP-seq comparisons.

https://cdn.elifesciences.org/articles/77595/elife-77595-fig3-data2-v3.xlsx
Figure 3—source data 3

Comparison of H3K27me2/3 ChIP-seq data and RNA-seq data in ∆iswi and ∆acf1.

https://cdn.elifesciences.org/articles/77595/elife-77595-fig3-data3-v3.xlsx

Loss of H3K27me2/3 or H3K36me3 in ∆iswi and ∆acf1 strains is not required for transcriptional upregulation of genes in facultative heterochromatin

We next asked if these changes in H3K27 or H3K36 methylation correlated with changes in gene expression in H3K27-methylated regions. We found that there was a negative correlation between H3K27me2/3 and gene expression in both ∆iswi (r=–0.404) and ∆acf1 (r=–0.36) strains (Figure 3E and F) while there was no correlation between H3K36me3 and gene expression in these strains (Figure 3G and H). Despite the negative correlation between H3K27me2/3 level and gene expression, the majority of upregulated genes—51% and 74% for ∆iswi and ∆acf1, respectively—had no significant loss of H3K27 methylation (Figure 3E and F). We looked at the distribution of H3K27me2/3 and H3K36me3 along the chromosomes (Figure 3I and J and Figure 3—figure supplement 2D) and found that while much of the H3K27me2/3 and H3K36me3 resembled wild type, large domains of H3K27me2/3 were lost in ∆iswi strains whereas more discrete decreases in H3K36me3 were observed in ∆iswi strains and to a lesser degree in ∆acf1 strains (Figure 3I–K).

We examined the gene expression changes along the left arm of LG IV in the ∆iswi strain and confirmed that many upregulated genes fell in regions that showed wild-type H3K27me2/3 and H3K36me3 (Figure 3K). ChIP-qPCR at the H3K27-methylated marker genes (NCU05173 and NCU07152) further validated the finding that loss of H3K27 methylation is not required for transcriptional upregulation in ∆iswi and ∆acf1 strains (Figure 3L). This is consistent with our previous findings showing that loss of H3K27 methylation is not a prerequisite for upregulating genes in facultative heterochromatin (Wiles et al., 2020). Taken together, these data show that iswi is required for normal H3K27me2/3 and H3K36me3, while loss of acf1 results in minor changes, suggesting that the ACF complex does not play a major role in directing or maintaining these histone modifications. Furthermore, loss of these histone marks is not required for transcriptional upregulation in facultative heterochromatin.

SET-7 promotes ACF1 association with facultative heterochromatin

To identify chromatin targets of the N. crassa ACF complex, we fused the Escherichia coli DNA adenine methyltransferase (van Steensel and Henikoff, 2000) to the C-terminus of endogenous ACF1 and assayed adenine methylated DNA fragments by sequencing (DamID-seq) (Zhou, 2012). We found that ACF1 localization is not restricted to one part of the genome, but rather appears to interact with chromatin genome-wide (Figure 4A and B). However, when set-7 was deleted, eliminating H3K27 methylation, ACF1 localization to H3K27-methylated regions was reduced relative to wild type, suggesting that H3K27 methylation, or SET-7 presence, promotes ACF1 interactions specifically with these genomic regions (Figure 4A and B, Figure 4—figure supplement 1A, B). These results were confirmed for two H3K27 methylation-marked regions (NCU05173 and Tel VIIL) by Southern hybridizations with genomic DNA from DamID experiments. In contrast, deletion of set-7 had no effect on ACF1-Dam localization at a euchromatic region (his-3) (Figure 4—figure supplement 1A). When we compared ACF1-Dam localization to that of a nonspecific control (Dam only; referred to as Free-Dam), we found that both constructs localized to non-H3K27-methylated genes at similar levels, and this was independent of set-7 presence (Figure 4C). In contrast, ACF1-Dam localized to H3K27-methylated genes more than Free-Dam and this increased localization was partially dependent on set-7 (Figure 4D). These data suggest that ACF1 association with facultative heterochromatin is promoted by, but not fully dependent on, an intact PRC2 complex and/or H3K27 methylation.

Figure 4 with 1 supplement see all
ACF1 localizes to H3K27me2/3-marked regions of the genome.

(A) Top track shows wild-type H3K27me2/3 levels based on ChIP-seq averaged from two biological replicates for one chromosome (LG III). Y-axis is 0–500 RPKM. Middle two tracks show DamID-seq average reads merged from two biological replicates for the indicated genotypes. Y-axis is 0–500 RPKM. Bottom track compares the DamID-seq reads from ∆set-7 strains to wild-type strains (shown above) displayed as the fold change between the two genotypes. Y-axis is –3–3. (B) Same as in (A), but showing an enlarged view of the right arm of LG III. Region shown is underlined in black in (A). (C) Average enrichment based on DamID-seq for each non-H3K27-methylated gene, scaled to 1 kb, ±500 base pairs, is plotted for the indicated strains. All lines represent average reads from two biological replicates except for Free-Dam which is from only one. TSS, transcription start site; TTS, transcription termination site. (D) Same as in (C), but for H3K27-methylated genes.

Loss of ACF has minor effects on nucleosome spacing

ACF-like complexes function differently in flies and yeast. In D. melanogaster, ACF acts globally to space nucleosomes evenly (Baldi et al., 2018), whereas in S. cerevisiae, the analogous Isw2 complex specifically moves the +1 nucleosome in the 5′ direction, toward the nucleosome-depleted region (NDR) (Whitehouse et al., 2007; Yen et al., 2012). To characterize nucleosome positioning in wild-type and mutant strains of N. crassa, we performed MNase digestion followed by high-throughput sequencing (MNase-seq). We first looked at nucleosome repeat length using the autocorrelation function (Braunschweig et al., 2009), which can analyze nucleosome positions independent of the TSS. When we looked genome-wide or considered only H3K27-methylated regions, we found only minor changes in nucleosome repeat length between wild-type and mutant strains (∆iswi, ∆acf1,iaf-3, ∆iaf-1, ∆iaf-2, and ∆set-7) (Figure 5—figure supplement 1A, B). This suggested that ISWI-containing complexes do not have major contributions to global nucleosome spacing in N. crassa or there is redundancy among these proteins.

Loss of ACF results in a downstream shift of the +1 nucleosome and transcriptional upregulation at a subset of H3K27-methylated genes

We next considered that the N. crassa ACF complex may function more like the S. cerevisiae Isw2 complex. For this analysis, we looked at nucleosome positions in the promoter region of genes that had regular nucleosome arrays (defined as spectral density SD; Baldi et al., 2018 score>2; n=7753) in at least one strain (wild type, ∆iswi, ∆acf1,iaf-3, ∆iaf-1, ∆iaf-2, or ∆set-7). We found that when all SD genes were considered, deletion of iswi or acf1 was more likely to result in a downstream shift (>30 bp) of the +1 nucleosome than when iaf-3, iaf-1, iaf-2, or set-7 were deleted (Figure 5—figure supplement 2A). This trend held when only H3K27-methylated SD genes (n=358) were considered (Figure 5A). Importantly, a significant portion of the H3K27-methylated genes with a shifted nucleosome is shared between iswi and acf1 (p<9.91E−13) (Figure 5B). These data suggest that ISWI and ACF1 may work in concert to position the +1 nucleosome at a subset of genes, including those in H3K27-methylated regions.

Figure 5 with 2 supplements see all
ISWI and ACF1 position the +1 nucleosome at H3K27-methylated, upregulated genes.

(A) Histogram of the number of H3K27-methylated SD genes (spectral density score for nucleosome order>2; n=358) that have the +1 nucleosome shifted downstream >30 base pairs when compared to wild type in the indicated mutant strains. Each point represents biological replicate 1, biological replicate 2, or analysis of the merged replicates and filled bar is the average of all three values. P values were determined with an unpaired t-test. (B) Venn diagram showing overlap of H3K27-methylated SD genes with a +1 nucleosome shifted downstream >30 bp when iswi or acf1 is deleted. P value was determined by hypergeometric probability test. (C–F) Average nucleosome signal at all H3K27-methylated SD genes plotted from MNase-seq data for the indicated mutants and wild type. The three colored lines represent biological replicate 1, biological replicate 2, and the average of the replicates for the strains indicated in the key. Arrows in (C) and (D) indicate the shifted +1 nucleosome. (G) Average nucleosome signal at SD genes that are upregulated (FDR <0.05) and marked by H3K27 methylation in ∆acf1 strains. The three colored lines represent biological replicate 1, biological replicate 2, and the average of the replicates. The boxed, shaded region is enlarged in the lower panel. (H) Same as panel (G), but for H3K27-methylated SD genes that are not upregulated in ∆acf1 strains. (I) Average nucleosome signal at SD genes that are upregulated (FDR <0.05) and marked by H3K27 methylation in ∆iswi strains. The three colored lines represent biological replicate 1, biological replicate 2, and the average of the replicates. Arrow indicates the shifted +1 nucleosome. (J) Same as (I), but for H3K27-methylated SD genes that are not upregulated in ∆iswi strains. (K, L) Average nucleosome signal at SD genes that are upregulated (FDR <0.05) and not marked by H3K27 methylation in ∆iswi (K) and ∆acf1 (L) strains. The three colored lines represent biological replicate 1, biological replicate 2, and the average of the replicates. (M) Same as (I), but for H3K27-methylated SD genes that are upregulated in ∆set-7.

Analysis of nucleosome positions at all SD genes in wild-type and mutant strains (∆iswi, ∆acf1,iaf-3, ∆iaf-1, ∆iaf-2, and ∆set-7) revealed some differences in occupancy at the –1 nucleosome but no global shift in nucleosome positions (Figure 5—figure supplement 2B). Because a genome-wide view can mask changes at specific targets and because the characteristic Isw2 5′ ‘pulling’ activity can only be appreciated when a subset of targets are examined (Yen et al., 2012; Ocampo et al., 2016; Donovan et al., 2021), we sought to limit our analysis to genes that might be targets of the ACF complex. Our inability to perform chromatin immunoprecipitation (ChIP) on ACF1 and limitations of DamID-seq precluded a strict analysis of direct ACF1 targets. Considering that our data support a functional role in transcriptional repression at H3K27-methylated genes, we restricted our analysis to these regions (H3K27-methylated SD genes; n=358). The MNase signal plots revealed that the +1 nucleosome shifted downstream in the absence of iswi or acf1 (Figure 5C and D); in contrast, no shift was seen when other ISWI-interacting partners (∆iaf-3, ∆iaf-1, and ∆iaf-2) were deleted (Figure 5E; Figure 5—figure supplement 2C, D). There was also no shift observed in ∆set-7 strains when all H3K27-methylated SD genes were considered (Figure 5F). These findings suggest that ISWI and ACF1 act to position the +1 nucleosome at a substantial subset of H3K27-methylated genes. Furthermore, SET-7, and hence H3K27 methylation, is not required for nucleosome positioning by ISWI/ACF1.

To test if the downstream nucleosome shift at H3K27-methylated genes in ∆iswi or ∆acf1 strains correlated with increased gene expression, we further focused our analysis to look at nucleosome positions in H3K27-methylated SD genes that were upregulated when iswi or acf1 was deleted. We found that the +1 nucleosome shifted 50-bp downstream on average at H3K27-methylated SD genes that were upregulated (FDR <0.05) in ∆acf1 strains (Figure 5G), whereas no such shift was seen in the +1 nucleosome of H3K27-methylated SD genes that were not upregulated in ∆acf1 strains (Figure 5H). Similarly, H3K27-methylated genes that were upregulated (FDR <0.05) in ∆iswi display a more prominent downstream shift of the +1 nucleosome than those genes that were not upregulated (Figure 5I and J). Taken together, these data suggest that positioning of the +1 nucleosome by ISWI and ACF1 at a subset of H3K27-methylated genes contributes to transcriptional repression.

To ensure that the nucleosome shifts observed at H3K27-methylated, upregulated genes in ∆iswi or ∆acf1 strains were not simply a consequence of the transcriptional activity, we looked at nucleosome positions in non-H3K27-methylated genes that are upregulated in these strains. We found no nucleosome shift at non-H3K27-methylated genes that are upregulated in ∆iswi or ∆acf1 (Figure 5K and L). This suggests that the upregulation at non-H3K27-methylated targets is through a different mechanism or is an indirect effect. We also looked at the nucleosome positions in H3K27-methylated genes that are upregulated in ∆set-7 (Figure 5M). We found no nucleosome shift at these genes. Taken together, this shows that transcriptional upregulation is not sufficient to induce a nucleosome shift. These findings support a model in which ACF acts directly at H3K27-methylated, upregulated genes.

ACF is a new player in the multifaceted repression of facultative heterochromatin

To gain a better understanding of how histone remodeling by ACF fits into our current framework of transcriptional repression by facultative heterochromatin, we compared gene expression profiles for ∆iswi and ∆acf1 to data sets for other genes that we identified as players in this repression: SET-7, the H3K27 methyltransferase; ASH1, an H3K36 methyltransferase; and EPR-1, an apparent H3K27 methyl-binding protein. We created a clustered heatmap of the gene expression data and found that five clusters emerged (Figure 6A). Cluster 1 included genes that were upregulated in every mutant strain except ∆epr-1. Cluster 2 contained genes that were highly upregulated in ∆iswi, ash1Y888F(a catalytic null) or ∆acf1 strains, while cluster 3 genes were only upregulated in ∆iswi or ash1Y888Fstrains. The fourth and largest cluster contained genes that were not highly upregulated in any mutant strain. The final cluster contained genes with more varied expression among the mutant strains, but perhaps showed some enrichment for genes upregulated in ∆set-7 and ∆epr-1 strains.

Figure 6 with 1 supplement see all
Multifaceted repression in facultative heterochromatin.

(A) Clustered heatmap made using mRNA-seq data for combined biological replicates of the indicated mutant strains. All H3K27-methylated genes that had reads in mRNA-seq data were included (n=821). Clusters (C1–C5) were determined by eye. (B) Model depicting our current framework of factors responsible for maintaining gene silencing in regions marked by H3K27 methylation. Loss of this methyl-mark itself is sufficient to activate a fraction of genes, in part because of loss of the H3K27 methyl-specific factor EPR-1. Repression of many other genes, in H3K27-methylated domains and elsewhere, depend on both H3K36 methylation by ASH1 and both components of the ACF complex (ISWI and ACF). Gray partial square represents an unknown H3K36 methyl binding protein. TF represents unknown transcription factor(s) that could recruit/direct the ACF complex.

The heatmap revealed striking overlap between gene expression profiles of ∆iswi and ash1Y888Fstrains, which prompted us to investigate this relationship further. We previously showed that iswi is not required for H3K36me2 (Figure 3—figure supplement 1), the predominant histone mark catalyzed by ASH1. We explored the possibility that ash1 is required for proper nucleosome positioning by the ACF complex. We performed MNase-seq in ash1 mutant strains and looked at nucleosome positions at H3K27-methylated genes that are upregulated when iswi is deleted. In contrast to the prominent downstream nucleosome shift we saw in this gene set in ∆iswi strains, we observed no nucleosome shift in ash1 mutant strains (Figure 6—figure supplement 1A). These data suggest that ISWI and ASH1 act in distinct, perhaps parallel, pathways to regulate a common set of H3K27-methylated genes.

Discussion

Control and function of facultative heterochromatin reflect the superimposition of a constellation of molecular mechanisms

Pioneering work on the Polycomb system of Drosophila revealed that methylation of lysine 27 of histone H3, catalyzed by Enhancer of Zeste in the PRC2 complex, is associated with, and important for, gene repression in facultative heterochromatin (Kassis et al., 2017). Although much has been learned, the importance of Polycomb repression in the development of multicellular organisms has stymied progress toward a full understanding of its control and function. Moreover, there are indications of variable underlying mechanisms. For example, the PRC1 complex, which is widely regarded as central to Polycomb function in Drosophila and higher organisms, is less conserved than PRC2, and at least in some organisms, is absent (Schuettengruber et al., 2017). Similarly, Polycomb response elements (PREs), cis-acting DNA sequences controlling the distribution of H3K27 methylation in Drosophila, do not appear to be universal (Kassis and Brown, 2013). The complexity and importance of the Polycomb system in multicellular organisms led us to dissect the control and function of H3K27 methylation in the filamentous fungus N. crassa. We defined the PRC2 complex of Neurospora, demonstrated that it methylates H3K27 in roughly 7% of the genome, and is necessary for repression of scores of genes even though it is not essential in this organism (Jamieson et al., 2013; McNaught et al., 2020). Utilization of special genetic resources for Neurospora revealed that the organism has at least two distinct forms of H3K27 methylation (Jamieson et al., 2018), namely: 1. position-dependent, associated with telomere regions and characterized by involvement of the Neurospora p55 homolog (NPF) and the PRC2 accessory subunit (PAS) (McNaught et al., 2020), and 2. position-independent, which is found interstitially and does not depend on NPF or PAS (Jamieson et al., 2018). We have also previously shown that ASH1, an H3K36 methyltransferase, is critical for maintaining repression of many genes, including most of those in facultative heterochromatin (Bicocca et al., 2018).

The nonessential nature of H3K27 methylation, and the convenience of Neurospora for genetics and biochemistry, allowed us to design and implement a powerful selection for mutants defective in silencing genes in facultative heterochromatin. This unbiased scheme revealed both expected factors required for repression, including members of the PRC2 complex, and unanticipated players such as EPR-1 (Wiles et al., 2020) and the ACF complex reported here. A particularly interesting general finding is that repression is not simply due to a linear pathway of factors. While some factors cooperate to maintain repression, results of mRNA-seq revealed considerable variation in the spheres of influence of the various factors. For example, loss of the H3K27 methyl-mark itself, or of the apparent H3K27 methyl-reader, EPR-1, each lead to derepression of somewhat different subsets of H3K27-methylated genes (Wiles et al., 2020), while loss of elements of the ACF remodeling machine leads to loss of silencing of a larger set of H3K27 methyl-marked genes, even without loss of this characteristic mark of facultative heterochromatin. The overall picture that is emerging is cartooned in Figure 6B with more specifics discussed below.

The N. crassa ACF complex is required for transcriptional repression at facultative heterochromatin

Our genetic identification of iswi and acf1 as genes required for silencing H3K27 methyl-marked loci is consistent with a growing body of evidence that chromatin structure plays a major role in the transcriptional status of genes. NDRs are characteristic of transcriptionally active promoters and are thought to allow access of the transcriptional machinery (Lai and Pugh, 2017). Conversely, nucleosomes can be positioned onto regulatory sequences in promoter regions by chromatin remodelers to cause repression (Whitehouse and Tsukiyama, 2006; Whitehouse et al., 2007). We found that nearly all H3K27-methylated genes that are upregulated in ∆acf1 also showed increased expression in ∆iswi, whereas ∆iswi had several uniquely upregulated genes. This is consistent with a model in which ACF1 is required for targeting ACF to chromatin targets but requires ISWI to catalyze nucleosome movement and allow for increased transcription. ISWI is also part of other protein complexes which may lead to direct or indirect upregulation of distinct genes.

ACF-like complexes are conserved from budding yeast (Tsukiyama et al., 1999) to humans (LeRoy et al., 2000), but most of the biochemical studies of these complexes have been done with yeast and flies, which, curiously, revealed apparent functional discrepancies. In yeast, Isw2 (homologous to the ACF complex of Drosophila) acts in promoter regions where it binds to the +1 nucleosome and moves it in the 5′ direction toward the NDR (Whitehouse et al., 2007; Yen et al., 2012; Kubik et al., 2019). In contrast, ACF in Drosophila has been characterized as a nonspecific nucleosome spacing and assembly factor promoting global chromatin regularity (Baldi et al., 2018). The distinct modes of action of ACF-like chromatin remodelers in yeast and Drosophila warrant further study in other organisms. Our investigation of nucleosome positioning activities of ISWI and ACF1 in N. crassa revealed that these factors are required for positioning the +1 nucleosome at a subset of genes, particularly those marked by H3K27 methylation. Thus, N. crassa ACF seems to function more like the S. cerevisiae Isw2 than the D. melanogaster ACF.

Although the detailed mechanism of recruitment and target selection for ACF-like complexes remains unclear, work in yeast implicates interactions of such complexes with transcription factors (Goldmark et al., 2000; Donovan et al., 2021). It was recently shown that the WAC domain of Itc1 in the Isw2 complex contains acidic residues required for binding to transcription factors and for nucleosome positioning at target promoters (Donovan et al., 2021). These residues (E33 and E40) are conserved in Neurospora (E32 and E39) but not in Drosophila, potentially accounting for the apparent less specific function of ACF in flies (Donovan et al., 2021). It will be of interest to determine if there are transcription factors that bind to facultative heterochromatin in N. crassa and mediate interactions with ACF1 to facilitate localization and activity of the ACF complex.

Our DamID-seq results are compatible with a ‘continuous sampling’ model proposed for some ISWI chromatin remodelers (Erdel et al., 2010). In this model, the ACF complex transiently interacts with chromatin (Gelbart et al., 2005) throughout the nucleus in an autoinhibited conformation (Clapier and Cairns, 2012; Ludwigsen et al., 2017) until some, still undefined, feature (Clapier and Cairns, 2012; Hwang et al., 2014; Ludwigsen et al., 2017; Donovan et al., 2021) releases the autoinhibition and allows it to engage, activate, and move nucleosomes by hydrolyzing ATP. Our results suggest that ACF localizes broadly throughout the genome but has specific activity at H3K27-methylated regions, raising the possibility that a feature of facultative heterochromatin influences ACF activity. The transient nature of this chromatin interaction could account for our inability to confirm our ACF1 DamID findings with ChIP. Attempts to identify the targets of the homologous complex by ChIP have been also unsuccessful in Drosophila (Scacchetti et al., 2018).

The N. crassa ACF complex positions the +1 nucleosome in promoters of H3K27-methylated genes to mediate transcriptional repression

In theory, nucleosome movement could be either a cause or consequence of transcriptional activation. Our finding that H3K27-methylated genes did not show changes in the position of the +1 nucleosome when they are derepressed by deletion of set-7 suggests that changes in nucleosome position are not simply due to transcriptional activation. Moreover, the fact that genes that are upregulated outside of facultative heterochromatin domains in strains with deletions of iswi or acf1 do not display a nucleosome shift suggests that ACF acts directly and has some specificity for H3K27-methylated regions. These findings support the idea that transcriptional derepression of H3K27-methylated genes in ∆iswi and ∆acf1 strains is a consequence of a misplaced +1 nucleosome. It is noteworthy that while Isw2-mediated repression is thought to occur by the placement of the +1 nucleosome over important DNA regulatory elements, occluding transcriptional machinery and/or general regulatory factors (Whitehouse et al., 2007; Yen et al., 2012), full repression at some targets, such as the early meiotic genes, also requires histone deacetylase activity from Rpd3 (Goldmark et al., 2000; Fazzio et al., 2001). Clearly, the mechanism of repression by ACF, including the identification of additional players, perhaps including transcription factors, histone deacetylases, and other chromatin modifying factors, deserves further study.

Conclusions

Despite differences in the modes of action of S. cerevisiae Isw2 and Drosophila ACF, their biological outcomes are the same—transcriptional repression (Goldmark et al., 2000; Fyodorov et al., 2004; Ocampo et al., 2016; Scacchetti et al., 2018). We found that nucleosome positioning by the N. crassa ACF complex also leads to transcriptional repression, particularly at H3K27-methylated regions of the genome, establishing the ACF complex as a player in transcriptional repression characteristic of facultative heterochromatin. It will be valuable to determine if interplay between Polycomb-mediated repression and ISWI chromatin remodelers holds in other organisms. Interestingly, ACF has been indirectly linked to Polycomb repression in flies (Scacchetti et al., 2018), and notably, ISWI components were identified in a screen for factors required for Polycomb repression in mammalian cells (Nishioka et al., 2018), raising the possibility that the role of the ACF complex in Neurospora is general.

Materials and methods

Key resources table
Reagent type (species) or resourceDesignationSource or referenceIdentifiersAdditional information
Strain, strain background (Neurospora crassa)MauricevilleFGSC 2225N51mat A; Mauriceville
Strain, strain background (N. crassa)Wild typeFGSC 2489N3752mat A; Oak Ridge
Strain, strain background (N. crassa)Wild typeFGSC 4200N3753mat a; Oak Ridge
Strain, strain background (N. crassa)Sad-1; his-3Wiles et al., 2020N3756mat A; Sad-1; his-3
Strain, strain background (N. crassa)set-7FGSC#11182N4718mat a; ∆set-7::hph
Strain, strain background (N. crassa)set-7Jamieson et al., 2018N4730mat A; ∆set-7::bar
Strain, strain background (N. crassa)ash1Y888FBicocca et al., 2018N4878mat A; his-3; ash1Y888F::3xFLAG::hph
Strain, strain background (N. crassa)pNCU07152::nat-1;set-7Wiles et al., 2020N5807Mat A; pNCU07152::nat-1;set-7::bar
Strain, strain background (N. crassa)pNCU07152::nat-1Wiles et al., 2020N5808mat a; pNCU07152::nat-1
Strain, strain background (N. crassa)iswiFGSC 11780N6170mat A; ∆iswi::hph
Strain, strain background (N. crassa)iswiThis studyN6171mat a; ∆iswi::hph
Strain, strain background (N. crassa)Mutant hunt strainWiles et al., 2020N6279mat a; pNCU05173::hph; pNCU07152::nat-1; his-3
Strain, strain background (N. crassa)iswiL430P original mutantThis studyN6606mat a; pNCU05173::hph; pNCU07152::nat-1; his-3; iswiL430P
Strain, strain background (N. crassa)pNCU07152::nat-1; ∆iswi::hphThis studyN6727mat a; pNCU07152::nat-1; ∆iswi::hph
Strain, strain background (N. crassa)ash1Y888FThis studyN6876mat a; ash1Y888F::3xFLAG::nat-1
Strain, strain background (N. crassa)ash1Y888FThis studyN6877mat a; ash1Y888F::3xFLAG::nat-1
Strain, strain background (N. crassa)EPR-1-DamWiles et al., 2020N7525mat A; epr-1::10xGly::Dam::nat-1
Strain, strain background (N. crassa)EPR-1-Dam; ∆eedWiles et al., 2020N7538mat a; epr-1::10xGly::Dam::nat-1;eed::hph
Strain, strain background (N. crassa)Free-Dam; ∆set-7This studyN7476mat A; ∆set-7::hph;his-3+::NLS(SV40)::Dam::3xFLAG::nat-1
Strain, strain background (N. crassa)Free-Dam; ∆set-7This studyN7477mat a; ∆set-7::hph;his-3+::NLS(SV40)::Dam::3xFLAG::nat-1
Strain, strain background (N. crassa)Free-DamThis studyN7802mat A; his-3+::NLS(SV40)::Dam::3xFLAG::nat-1
Strain, strain background (N. crassa)iswiL430P complement-ation strainThis studyN7810mat a; pNCU05173::hph; pNCU07152::nat-1; his-3+::Pccg-1::3xFLAG::iswiWT; iswiL430P
Strain, strain background (N. crassa)pNCU07152::nat-1;iaf-2This studyN7941mat a; pNCU07152::nat-1;iaf-2::hph
Strain, strain background (N. crassa)acf1D161fs original mutantThis studyN7953mat a; pNCU05173::hph; pNCU07152::nat-1; his-3; acf1D161fs
Strain, strain background (N. crassa)pNCU07152::nat-1;acf1This studyN7956mat a; pNCU07152::nat-1;mus-52::baracf1::hph
Strain, strain background (N. crassa)pNCU07152::nat-1;iaf-3::hphThis studyN7960mat A; pNCU07152::nat-1;iaf-3::hph
Strain, strain background (N. crassa)pNCU07152::nat-1;iaf-1:hphThis studyN7961mat a; pNCU07152::nat-1;iaf-1:hph
Strain, strain background (N. crassa)iaf-3This studyN7966mat A; ∆iaf-3::hph
Strain, strain background (N. crassa)ACF1-HAThis studyN7971mat a; ∆mus-52::bar acf1::HA::hph
Strain, strain background (N. crassa)IAF-1-HAThis studyN7973mat A; ∆mus-52::bar; iaf-1::HA::hph
Strain, strain background (N. crassa)∆iaf-2This studyN7988mat a; ∆iaf-2::hph
Strain, strain background (N. crassa)∆iaf-2This studyN7989mat a; ∆iaf-2::hph
Strain, strain background (N. crassa)∆iaf-1FGSC 12715N7990mat a; ∆iaf-1::hph
Strain, strain background (N. crassa)∆iaf-1This studyN7992mat a; ∆iaf-1::hph
Strain, strain background (N. crassa)acf1This studyN8016mat a; ∆acf1::hph
Strain, strain background (N. crassa)acf1This studyN8017mat a; ∆acf1::hph
Strain, strain background (N. crassa)∆iaf-3This studyN8018mat A; ∆iaf-3::hph
Strain, strain background (N. crassa)IAF-3-HAThis studyN8071mat A; pNCU07152::nat-1; iaf-3::HA::hph
Strain, strain background (N. crassa)IAF-2-HAThis studyN8075mat a; pNCU07152::nat-1; iaf-2::HA::hph
Strain, strain background (N. crassa)ACF1-Dam; ∆set-7This studyN8113mat A;set-7::hph; ∆mus-52::bar acf1::Dam::nat-1
Strain, strain background (N. crassa)ACF1-Dam; ∆set-7This studyN8114mat a;set-7::hph; ∆mus-52::bar acf1::Dam::nat-1
Strain, strain background (N. crassa)ACF1-DamThis studyN8115mat A; ∆mus-52:bar acf1::Dam::nat-1
Strain, strain background (N. crassa)acf1D161fscomplement-ation strainThis studyN8142mat a; pNCU05173::hph; pNCU07152::nat-1; his-3+::Pccg-1::acf1WT::mCherry; acf1D161fs
Strain, strain background (N. crassa)ACF1-DamThis studyN8146mat a; ∆mus-52::bar acf1::Dam::nat-1
Strain, strain background (N. crassa)pNCU07152::nat-1;hfp-1This studyN8197mat a; pNCU07152::nat-1;hfp-1::hph
Sequence-based reagenthH4_qPCR_FP(4082)Jamieson et al., 2013ChIP-qPCR primerCATCAAGGGGTCATTCAC
Sequence-based reagenthH4_qPCR_RP (4083)Jamieson et al., 2013ChIP-qPCR primerTTTGGAATCACCCTCCAG
Sequence-based reagentNCU07152_promoter_FP(6565)Wiles et al., 2020ChIP-qPCR primerCGGTTCCAAAACTGCCCCTGTG
Sequence-based reagentNCU07152_promoter_RP(6645)Wiles et al., 2020ChIP-qPCR primerCTCAGCGGGGTATATCAACGGC
Sequence-based reagentNCU05173_promoter_FP(6567)Wiles et al., 2020ChIP-qPCR primerGCATTACCCTCGACAGGGTCTG
Sequence-based reagentNCU05173_promoter_RP(6646)Wiles et al., 2020ChIP-qPCR primerGCTACCACCATGTGAAGCTCTGG
Sequence-based reagenthis-3_FP(1665)Klocko et al., 2019Southern probe primersGACGGGGTAGCTTGGCCCTAATTAACC
Sequence-based reagenthis-3_RP(3128)Klocko et al., 2019Southern probe primersCGATTTAGGTGACACTATAG
Sequence-based reagentTel_VIIL_FP(5271)Wiles et al., 2020Southern probe primersGGCATCCGTGGGTGTCCCAG
Sequence-based reagentTel_VIIL_RP(5272)Wiles et al., 2020Southern probe primersTTCCCGTCCCTACCAGGCAT
Sequence-based reagentNCU05173_FP(6567)Wiles et al., 2020Southern probe primersGCATTACCCTCGACAGGGTCTG
Sequence-based reagentNCU05173_RP(6568)Wiles et al., 2020Southern probe primersCCTGTTCGAGTTATCGGTGTTG
Antibodyα-H3K27me2/3 (mouse monoclonal)Active MotifCat. #39536Chromatin immunoprecipitation(2 µl ChIP-seq; 3 µl ChIP-qPCR)
Antibodyα-H3K36me2(rabbit polyclonal)AbcamCat. #ab9049Chromatin immunoprecipitation(2 µl)
Antibodyα-H3K36me3(rabbit polyclonal)AbcamCat. #ab9050Chromatin immunoprecipitation(2 µl)
Antibodyα-HA(mouse monoclonal)MBLCat. #180-3Immunoprecipitation(20 µl)
Antibodyα−FLAG M2 affinity gel (mouse monoclonal)Sigma-AldrichCat. #A2220Immunoprecipitation(400 µl)
Peptide, recombinant proteinHA peptideThermo Fisher ScientificCat. #26184Elution
Peptide, recombinant protein3× Flag peptideAPExBIOCat. #A6001Elution

Strains, media, and growth conditions

Request a detailed protocol

All N. crassa strains were grown as previously described (Wiles et al., 2020) and are listed in the Key resources table. Technical replicates are defined as experimental repeats with the same strain. Biological replicates are defined as experiments performed using a different strain with the same genotype.

Selection for mutants defective in Polycomb silencing

Request a detailed protocol

The selection was carried out as previously described (Wiles et al., 2020). Briefly, conidia from strain N6279 were mutagenized with UV radiation and subjected to selection with Hygromycin B or Nourseothricin. Resistant colonies were grown and crossed to strain N3756 to generate homokaryons.

Whole-genome sequencing, mapping, and identification of mutants

Request a detailed protocol

Whole-genome sequencing, SNP mapping, and identification of mutants were performed as previously described (Wiles et al., 2020). Briefly, antibiotic-resistant, homokaryotic mutants were crossed to a genetically polymorphic Mauriceville strain and approximately 15–20 antibiotic-resistant progeny were pooled and prepared for whole-genome sequencing using the Nextera Kit (Illumina, FC-121-1030). Mapping of the critical mutations was performed as previously described (Hunter, 2007; Pomraning et al., 2011). FreeBayes and VCFtools were used to identify novel genetic variants present in pooled mutant genomic DNA (Danecek et al., 2011; Garrison and Marth, 2012). All whole-genome sequencing data are available on NCBI Sequence Reads Archive (PRJNA714693).

Immunoprecipitation followed by MS

Request a detailed protocol

Strains N7810 (his-3::Pccg::3xFLAG-iswi), N7971 (endogenous acf1-HA), N8071 (endogenous iaf-3-HA), N7973 (endogenous iaf-1-HA), and N8075 (endogenous iaf-2-HA) were grown and protein extracted as previously described (McNaught et al., 2020) except that 500ml cultures were used. Purification of 3×FLAG-tagged protein was performed as previously described (McNaught et al., 2020). For HA-tagged proteins, the same procedure was used except that 20 µg of α-HA antibody (MBL 180-3) was bound to 400-µl equilibrated Protein A agarose (Invitrogen, 15918014) by rotating at room temperature for 1 hr and washed 3× with extraction buffer and protein was eluted 3× with 300µl of 1 mg/ml HA peptide (Thermo Fisher Scientific, 26184) in 1× TBS. Samples were sent to and processed by the UC Davis Proteomics Core Facility for MS and analysis.

RNA isolation, RT-qPCR, and mRNA-seq

Request a detailed protocol

Total RNA was extracted from germinated conidia as previously described (Wiles et al., 2020) and used for mRNA-seq library preparation (Klocko et al., 2016). Sequencing was performed by the University of Oregon Genomics and Cell Characterization Core Facility.

mRNA-seq data analysis

Request a detailed protocol

Sequence reads were aligned to the N. crassa genome (OR74A) using STAR program (version 2.7.3a). Total aligned reads per N. crassa gene were calculated using RSEM software (version 1.3.1) and normalized using DESeq2 software (version 1.24.0). Batch effects were corrected using R package, limma (version 3.44.1). FDR <0.05 and abs(log2 fold change)>2 were used as a threshold to identify significantly up- or downregulated genes.Clustered heatmaps were generated using all H3K27-methylated genes that had reads in the mRNA-seq data sets (n=821). Genes were sorted by giving the highest priority to common upregulated genes. If the number of common gene sets was the same, they were prioritized by genes upregulated in the following order: ∆iswi, ash1Y888F, ∆acf1, ∆set-7, and ∆epr-1. Genes were further sorted by log2 fold change value. All sequencing files are available on the NCBI GEO database (GSE168277).

ChIP, ChIP-qPCR, and ChIP-seq

Request a detailed protocol

H3K27me2/3 ChIP using α-H3K27me2/3 antibody (Active Motif, 39536), which recognizes di- or trimethylated H3K27, was performed as previously described (Wiles et al., 2020). H3K36me2 and H3K36me3 ChIP using α-H3K36me2 (Abcam, ab9049) and α-H3K36me3 (Abcam, ab9050) antibodies were performed as previously described (Bicocca et al., 2018). The isolated DNA was used for qPCR (see Key resources table for primers) or prepared for sequencing (Wiles et al., 2020). Sequencing was performed by the University of Oregon Genomics and Cell Characterization Core Facility.

ChIP-seq data analysis

Request a detailed protocol

Mapping, visualization, and analysis of ChIP-sequencing reads was performed as previously described (Wiles et al., 2020). H3K27me2/3 ChIP-seq tracks were normalized using RPKM and H3K36me3 tracks were normalized to 10 million reads using HOMER (Heinz et al., 2010). To generate scatter plots for H3K27me2/3, H3K36me2, and H3K36me3 in ∆iswi and ∆acf1 strains ChIP-seq normalized scores were calculated using HOMER to normalize the total tag to 10 M. Bam files from replicates were merged using ‘samtools,’ then normalized using HOMER to make mixed data. ‘bigWigAverageOverBed’ from kentUtils was used to generate the average at each gene. Differentially enriched genes (p-value <0.05 and log2 fold change>1 for gains and <1 for losses) were defined using edgeR package of R (version 3.30.3). Average scores from two replicates were used for the analysis. Average scores from the mix were used for the scatter plot. All sequencing files are available on the NCBI GEO database (GSE168277).

DamID Southern hybridization and sequencing

Request a detailed protocol

Southern hybridization was carried out as previously described (Miao et al., 2000) with probes generated by PCR amplification (see Key resources table for primers) from wild-type N. crassa genomic DNA (NCU05173, TelVIIL) or plasmid pBM61 (his-3). Genomic DNA was prepared for DamID-seq as previously described (Zhou, 2012) with the modifications we have reported (Wiles et al., 2020). Sequencing was performed by the University of Oregon Genomics and Cell Characterization Core Facility. The ‘Free Dam’ strain had an N-terminal NLS (SV40) and a C-terminal 3× FLAG tag and was expressed from the his-3 locus.

DamID-seq data analysis

Request a detailed protocol

DamID-seq mapping and analysis were done using the Galaxy public server (Afgan et al., 2018). The Barcode Splitter was used to filter for reads with a GATC at the 5′ end and these reads were mapped using Bowtie2 (Langmead and Salzberg, 2012). Files for biological replicates were merged using MergeBam. Merged bam files were used as input for bamCoverage (RPKM, 50-bp bins) to generate bigwig files for viewing on IGV and running bigwigCompare. The output from bamCoverage was used with computeMatrix to generate files to use for plotProfile and output graphs. All sequencing files are available on the NCBI GEO database (GSE168277).

MNase digestion and sequencing

Request a detailed protocol

N. crassa cells were grown and digested with micrococcal nuclease as previously described (McKnight et al., 2021) with the following modifications. MNase (Takara) concentration was optimized for each strain to yield ~80%–90% mononucleosomes (20 units for N3752, N3753, N7966, N8018, N7990, N7992, N7988, and N7989; 40 units for N6877; 60 units for N4718; and 80 units for N4730, N6170, N6171, N6876, N8016, and N8017). All digestions were for 10 min at 37°C, RNase (40 µg) treatment was for 1.5 h at 42°C, and proteinase K (200 µg) treatment was for 1 hr at 65°C. About 10 µg of gel-purified mononucleosome DNA was prepared for high-throughput sequencing using the NEBNext DNA Library Prep Master Mix Set for Illumina (NEB). Sequencing was performed by the University of Oregon Genomics and Cell Characterization Core Facility.

MNase-seq data analysis

Request a detailed protocol

Paired-end sequence reads were aligned to the N. crassa genome (OR74A) using Bowtie2 (version 2.3.3) with the option ‘-q -p 4X 250 --no-discordant --no-mixed --no-unal.’ Paired-end alignment reads with maximum 250-bp distance gap between them were used in subsequent analysis. This length corresponds to mononucleosomes. Only correctly aligned paired-end alignment reads were filtered using samtools (version 1.5) commands ‘samtools view –hf 0x2 input.bam | grep –v “XS:i:”’ Dyad Coverage was calculated using the scripts (03_PNA_SDE.R) (Baldi et al., 2018). All sequencing files are available on the NCBI GEO database (GSE168277).

Spectral density estimation

Request a detailed protocol

The spectral density (SD) score corresponding to periods of 182 bp was calculated using the scripts (cov2spec.R) (Baldi et al., 2018). SD score was normalized as Z-score: (log2(SD score)−average)/standard deviation. Regions with the average Z-score threshold of 2 were defined as the domain with a regular nucleosome array.

Autocorrelation function

Request a detailed protocol

The autocorrelation function (Braunschweig et al., 2009) was calculated for the dyad coverage vectors for the lag length of 1000 bp. Nucleosome repeat lengths were obtained by linear regression of the first and second autocorrelation peak positions with zero intercept. The slope of the regression was defined as repeat length.

Estimation of +1 nucleosome position

Request a detailed protocol

The average score of dyad coverage vector for every 182 bp using the region –100 bp to +1000 bp from TSS was calculated for each gene. The closest peak from TSS was defined as +1 nucleosome position.

Data availability

All RNA-seq, ChIP-seq, DamID-seq and MNase-seq data generated in this study have been submitted to the NCBI Gene Expression Omnibus (GEO; https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE168277. All whole genome sequencing data haven been submitted to the NCBI Sequence Read Archive (SRA, https://www.ncbi.nlm.nih.gov/sra) under accession number PRJNA714693.

The following data sets were generated
    1. Selker EU
    (2022) NCBI Gene Expression Omnibus
    ID GSE168277. The ACF chromatin remodeling complex is essential for Polycomb repression.
    1. Selker EU
    (2022) NCBI BioProject
    ID PRJNA714693. The ACF chromatin remodeling complex is essential for Polycomb repression.

References

  1. Book
    1. Zhou V.
    (2012)
    Methods for Global Characterization of Chromatin Regulators in Human Cells
    Harvard University Press.

Decision letter

  1. Jerry L Workman
    Reviewing Editor; Stowers Institute for Medical Research, United States
  2. Kevin Struhl
    Senior Editor; Harvard Medical School, United States
  3. Jerry L Workman
    Reviewer; Stowers Institute for Medical Research, United States

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Decision letter after peer review:

[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]

Thank you for submitting the paper "The ACF Chromatin Remodeling Complex is Essential for Polycomb Repression" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, including Jerry L Workman as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by a Senior Editor.

The three reviewers discussed the manuscript in depth. The reviewers all felt that the manuscript is interesting and contains novel findings. Other aspects of the manuscript seemed somewhat underdeveloped and needed additional experimentation. In particular, the reviewers wanted to see the effect of the mutants on H3K36 methylation, more evidence that nucleosome repositioning was responsible for derepression and the dependence of nucleosome positioning by ISWI on H3K27 methylation. As it is anticipated that these experiments will require more time than allowed per eLife policy for revision, we will not further consider this version of the manuscript, but we would be to consider a new version if the reviewers' comments can be addressed.

Reviewer #1:

Polycomb repression of heterochromatic genes differs in different organisms but has been most widely studied in Drosophila. Neurospora lacks components of Drosophila polycomb repression complexes.

Using a powerful forward genetic screen the authors found that components of the ACF complex were required to maintain repression of H3K27 methylated heterochromatic genes in Neurospora.

ACF binds widely to chromatin across the genome and is not restricted to heterochromatic genes. This indicates that it also functions outside of heterochromatin. Its interaction with heterochromatin is affected somewhat by the loss of H3K27 methylation.

ACF appears to be necessary to position the +1 nucleosome over the promoter of H3K27 methylated heterochromatic genes.

1. "We found that while the majority of gene expression changes observed upon loss of ISWI or ACF1 occurred outside of H3K27-methylated domains" Do they authors suggest that ACF performs a specialized function on H3K27 methylated domains or that it does the same function on all genes with a different outcome on H3K27 methylated genes? Please clarify.

2. A table presenting at least the relative parts of the mass spec data needs to be included in the manuscript so the readers can judge the spectrum counts relevance.

3. Figure 3A is a strange representation of ms data and probably should be removed. The table above should present that data.

4. The authors should limit their conclusions about Iswi protein complexes etc. Without more biochemistry we don't know how many complexes there are, and which components are in each. This would require co-fractionation experiments and probably more purifications. The authors can conclude that the interactions they detect by ms are consistent with iswi being in multiple complexes similar to that described in other organisms.

Reviewer #2:

In this paper Wiles et al. show that mutations in the iswi and acf genes, which encode components of a nucleosome remodeling complex, lead to expression of a subset of H3K27me-repressed genes. The strengths of the paper include the detailed genomic analysis supporting the statements that Iswi and Acf regulate a subset of H3K27me3-repressed genes. Data showing that the +1 nucleosome shifts 50bp in H3K27me-genes upregulated in the iswi mutant is also very strong. There is strong data documenting the proteins that Iswi interacts with in N. crassa. The data showing the nucleosome shift in the acf mutant is not as strong. The summary figure is highly speculative because there is no data for discrete localization of Acf. Another piece of data that is lacking is what happens to H3K36me in iswi and acf mutants. Knowing this is important because a similar set of genes seem to be derepressed in an ash1 mutant as in the acf and iswi mutants, although the level of depression in ash1 is not as great as in iswi mutants. The summary diagram shows loss of H3K36me as a separate mechanism than loss of the ACF complex. We don't know that since there was no analysis of H3K36me in iswi or acf mutants. Still, the major findings of the paper are important.

1) The paper would be strengthened by an examination of H3K36me in an iswi mutant.

2) Please comment more extensively on why the iswi mutant phenotype is so much stronger than the acf phenotype.

3) How do you think Acf gets localized to H3K27me domains?

4) In Figure 4 you show by qPCR that H3K27me is still present at the two genes used in the initial mutant screen. Please show additional examples of the H3K27me levels on other upregulated genes in the iswi and acf mutants. Are the H3K27me levels altered in any of these genes, or did I miss this?

Reviewer #3:

This manuscript clearly shows that, in Neurospora, ACF chromatin remodeler represses some of H3K27-methylated genes in an H3K27me-dependent fashion. This is the first report to demonstrate ISWI class remodeler functions through H3K27me. The main weakness of the manuscript is the modest impact on mechanistic understanding of how ACF or H3K27me functions. Also, the authors' model needs additional support to firmly establish the causality between ACF-dependent nucleosome repositioning and transcriptional repression.

In this manuscript, the authors showed that the ACF chromatin remodeler in N. crassa is required for repression of H3K27-methylated genes. The authors further demonstrated that ACF is targeted to these genes through H3K27me, and proposed a model in which ACF-dependent nucleosome repositioning around TSSs causes transcriptional repression. Molecular mechanisms underlying transcriptional repression at facultative heterochromatin is an important topic, and the authors provided convincing data that ACF plays critical roles through H3K27me, which has not been directly demonstrated for any ISWI class remodelers. On the other hand, this manuscript does not provide much advancement on molecular mechanism. Even if the authors' model that ACF-dependent nucleosome repositioning causes transcriptional repression is correct (see below), that mechanism has been described in budding yeast. Therefore, the mechanistic novelty is the H3K27-dependent ACF targeting.

Another issue in the authors' model that ACF-dependent nucleosome repositioning is the cause of transcriptional derepression. As far as I can tell, the data are also equally consistent with the possibility that ACF loss causes transcriptional derepression through other means, and the nucleosomes around TSSs reposition as a result of transcriptional derepression. The authors need to establish the causality more firmly. For example, are there loci where ACF repositions nucleosomes around TSSs similarly to other targets, but transcription is not derepressed upon ACF mutations? If there are many genes that fall into category, that would be consistent with the authors' model. On the other hand, if the authors look at genes that are not direct ACF targets but are derepressed similarly to ACF targets upon ACF mutations, do they see nucleosome repositioning similar to ACF targets? If not, that would argue transcriptional derepression is not sufficient to reposition nucleosomes around TSSs, supporting the authors' model.

1. Line 122: What is an "early condition phenotype"?

2. Lines 233-238: What is the conclusion drawn from this data?

3. Figure 2: It is interesting that transcriptional de-repression is much more robust in iswi mutant than acf1. Given that ISWI purification did not yield other form of ISWI complexes, this is not because of other forms of ISWI-containing remodelers (CHRAC also depends on ACF1). So, another possibility is that, in an iswi mutant, ACF1 functions in a dominant negative fashion to derepress genes further. If this is the case, iswi acf1 double mutants should behave like an acf1 single mutant. The authors should test this for a few targets genes to establish how ACF functions.

4. Figure 2S1: ~1/3 of genes repressed by Set7 are not repressed by ACF. Do these genes have any special features?

5. Do ACF signals by Dam-ID mapping show something specific at H3K27me genes that are derepressed and/or show changes in nucleosome positioning? If the authors can show specific localization at these genes, it would help support the model that the transcriptional effects mediated by ACF are direct.

6. ISWI regulates only a small subset of H3K27me genes (98 of 836). Are there unique properties of these genes that may contribute to our understanding of their mechanism of derepression? For example, are they regulated by a common transcription factor? Are their nucleosome positions in WT different from H3K27me genes not regulated by ISWI?

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for submitting your article "The ACF chromatin remodeling complex is essential for Polycomb repression" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, including Jerry L Workman as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Kevin Struhl as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Jerry L Workman (Reviewer #1).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

In this re-submitted manuscript, the authors addressed issues raised by the reviewers. As a result, this manuscript was substantially strengthened. The main conclusions from Figures1 through 3 are consistent with those of Kamei et al. (2021). However, this manuscript provides strong evidence that ACF directly functions to promote Polycomb-dependent repression through chromatin remodeling, which has not been demonstrated. In addition, PRC2/H3K27me-dependent ACF targeting is novel. Finally, the authors' model that facultative chromatin can be classified into several groups based on their dependence on SET7, ASH1, and ACF (Figure 6) is potentially important for guiding future research directions of the field.

The data that the effect of ACF on the position of +1 nucleosomes is distinct between H3K27me positive vs negative loci (Figure 5, Figure 5 Supplemental 2) is quite interesting. Is the degree of targeting of ACF and its effect on transcription similar between H3K27me positive and negative genes? If so, this result suggests currently unknown effects of H3K27me on ACF activity. In this case, Figure 5 Supplemental 2 should be move to the main Figure and the result deserves more extensive discussions as this would increase the impact of the manuscript. On the other hand, if the degree of ACF targeting and/or the level of transcription can explain the difference in the +1 nucleosome position, it needs to be stated in the text.

https://doi.org/10.7554/eLife.77595.sa1

Author response

[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]

Reviewer #1:

Polycomb repression of heterochromatic genes differs in different organisms but has been most widely studied in Drosophila. Neurospora lacks components of Drosophila polycomb repression complexes.

Using a powerful forward genetic screen the authors found that components of the ACF complex were required to maintain repression of H3K27 methylated heterochromatic genes in Neurospora.

ACF binds widely to chromatin across the genome and is not restricted to heterochromatic genes. This indicates that it also functions outside of heterochromatin. Its interaction with heterochromatin is affected somewhat by the loss of H3K27 methylation.

ACF appears to be necessary to position the +1 nucleosome over the promoter of H3K27 methylated heterochromatic genes.

1. "We found that while the majority of gene expression changes observed upon loss of ISWI or ACF1 occurred outside of H3K27-methylated domains" Do they authors suggest that ACF performs a specialized function on H3K27 methylated domains or that it does the same function on all genes with a different outcome on H3K27 methylated genes? Please clarify.

To address this question, we looked at the nucleosome positions at genes that are upregulated in ∆iswi and ∆acf1 but are not H3K27me (for SD genes that are included in nucleosome shift analysis: n=335; n=146, respectively). We found that there was no observable shift in the +1 nucleosome at these genes. This suggests that ACF is performing a specialized function (moving the +1 nucleosome) at H3K27me genes and that the upregulation seen in these mutants at non-H3K27me genes is either via a different mechanism or are indirect effects. It is also possible that there are other direct targets of ACF, but we are not able to see the nucleosome shift when we look at all non-H3K27me, upregulated genes. We have added two figure panels as well as text to clarify this point (Figure 5 —figure supplement 2C,D).

2. A table presenting at least the relative parts of the mass spec data needs to be included in the manuscript so the readers can judge the spectrum counts relevance.

We have included a table with spectrum counts for relevant proteins as Figure 2 —figure supplement 1 and all mass spec data can be found as source data files for Figure 2.

3. Figure 3A is a strange representation of ms data and probably should be removed. The table above should present that data.

See above. We have also kept this representation as we feel is includes some information that is lacking in the table.

4. The authors should limit their conclusions about Iswi protein complexes etc. Without more biochemistry we don't know how many complexes there are, and which components are in each. This would require co-fractionation experiments and probably more purifications. The authors can conclude that the interactions they detect by ms are consistent with iswi being in multiple complexes similar to that described in other organisms.

We have limited our conclusions regarding the components in putative ISWI complexes. We deleted the following “were also identified in our ISWI pull-downs, albeit below our threshold, and suggest the presence of an N. crassa CHRAC complex.”

Reviewer #2:

In this paper Wiles et al. show that mutations in the iswi and acf genes, which encode components of a nucleosome remodeling complex, lead to expression of a subset of H3K27me-repressed genes. The strengths of the paper include the detailed genomic analysis supporting the statements that Iswi and Acf regulate a subset of H3K27me3-repressed genes. Data showing that the +1 nucleosome shifts 50bp in H3K27me-genes upregulated in the iswi mutant is also very strong. There is strong data documenting the proteins that Iswi interacts with in N. crassa. The data showing the nucleosome shift in the acf mutant is not as strong. The summary figure is highly speculative because there is no data for discrete localization of Acf. Another piece of data that is lacking is what happens to H3K36me in iswi and acf mutants. Knowing this is important because a similar set of genes seem to be derepressed in an ash1 mutant as in the acf and iswi mutants, although the level of depression in ash1 is not as great as in iswi mutants. The summary diagram shows loss of H3K36me as a separate mechanism than loss of the ACF complex. We don't know that since there was no analysis of H3K36me in iswi or acf mutants. Still, the major findings of the paper are important.

We feel that the data showing the nucleosome shift in ∆acf1 are quite strong (Figure 5I). We agree that the model is somewhat speculative but we feel it is useful and have addressed concerns e.g. by performing H3K36me ChIP-seq (more below). Although we were unable to ChIP ACF, we find the combination of the ACF1-DamID and nucleosome shifts specifically at the K27me upregulated genes in ∆acf1 provides good evidence that ACF is acting at and localizing to these genomic locations. We have added H3K36me ChIP data in ∆acf1 and ∆iswi (more below).

1) The paper would be strengthened by an examination of H3K36me in an iswi mutant.

We have added H3K36me2 and H3K36me3 ChIP-seq data for ∆iswi and ∆acf1 to Figure 3. Briefly, we found no changes in H3K36me2 and some loss of H3K36me3 in these strains.

2) Please comment more extensively on why the iswi mutant phenotype is so much stronger than the acf phenotype.

We are not certain why the upregulation at H3K27me2/3 marked genes is higher in ∆iswi than in ∆acf1. We presume that the iswi mutant upregulates more genes than acf1 because ISWI is part of other protein complexes thus leading to indirect effects. To clarify, we have added the following sentences to the discussion.

“We found that nearly all H3K27 methylated genes that are upregulated in ∆acf1 also showed increased expression in ∆iswi, whereas ∆iswi had several uniquely upregulated genes. This is consistent with a model in which ACF1 is required for targeting ACF to chromatin targets but requires ISWI to catalyze nucleosome movement and allow for increased transcription. ISWI is also part of other protein complexes which may lead to direct or indirect upregulation of distinct genes.”

3) How do you think Acf gets localized to H3K27me domains?

ACF may interact with transcription factors to target H3K27me domains. Targeting of ACF by transcription factors been demonstrated in budding yeast (although not H3K27me targeting as this mark is absent in yeast) and is discussed in the Discussion section.

4) In Figure 4 you show by qPCR that H3K27me is still present at the two genes used in the initial mutant screen. Please show additional examples of the H3K27me levels on other upregulated genes in the iswi and acf mutants. Are the H3K27me levels altered in any of these genes, or did I miss this?

We used the RNA-seq and ChIP-seq data to address this question and added two new figure panels (Figure 3E,F) showing a scatter plot of the correlation between gene expression changes and changes in H3K27me2/3 level at all H3K27me2/3-marked genes. We found a substantial portion of genes that are upregulated in ∆iswi and ∆acf1 do not show significant loss of H3K27me2/3. We also added expression data for ∆iswi with corresponding ChIP-seq tracks in Figure 3K.

Reviewer #3:

This manuscript clearly shows that, in Neurospora, ACF chromatin remodeler represses some of H3K27-methylated genes in an H3K27me-dependent fashion. This is the first report to demonstrate ISWI class remodeler functions through H3K27me. The main weakness of the manuscript is the modest impact on mechanistic understanding of how ACF or H3K27me functions. Also, the authors' model needs additional support to firmly establish the causality between ACF-dependent nucleosome repositioning and transcriptional repression.

In this manuscript, the authors showed that the ACF chromatin remodeler in N. crassa is required for repression of H3K27-methylated genes. The authors further demonstrated that ACF is targeted to these genes through H3K27me, and proposed a model in which ACF-dependent nucleosome repositioning around TSSs causes transcriptional repression. Molecular mechanisms underlying transcriptional repression at facultative heterochromatin is an important topic, and the authors provided convincing data that ACF plays critical roles through H3K27me, which has not been directly demonstrated for any ISWI class remodelers. On the other hand, this manuscript does not provide much advancement on molecular mechanism. Even if the authors' model that ACF-dependent nucleosome repositioning causes transcriptional repression is correct (see below), that mechanism has been described in budding yeast. Therefore, the mechanistic novelty is the H3K27-dependent ACF targeting.

Another issue in the authors' model that ACF-dependent nucleosome repositioning is the cause of transcriptional derepression. As far as I can tell, the data are also equally consistent with the possibility that ACF loss causes transcriptional derepression through other means, and the nucleosomes around TSSs reposition as a result of transcriptional derepression. The authors need to establish the causality more firmly. For example, are there loci where ACF repositions nucleosomes around TSSs similarly to other targets, but transcription is not derepressed upon ACF mutations? If there are many genes that fall into category, that would be consistent with the authors' model. On the other hand, if the authors look at genes that are not direct ACF targets but are derepressed similarly to ACF targets upon ACF mutations, do they see nucleosome repositioning similar to ACF targets? If not, that would argue transcriptional derepression is not sufficient to reposition nucleosomes around TSSs, supporting the authors' model.

We feel that our finding that H3K27-methylated genes that are upregulated when the H3K27 methyltransferase set-7 is deleted do not have a nucleosome shift suggest that transcriptional upregulation in regions of facultative heterochromatin is not sufficient to result in the +1 nucleosome shift (we have moved this figure panel from the supplementary figures to the main figures – Figure 5M). We have also added new data (Figure 5 —figure supplement 2C,D) showing that non-H3K27 methylated genes that are upregulated in ∆iswi and ∆acf1 do not have a nucleosome shift, again suggesting that transcriptional upregulation is not sufficient to induce a nucleosome shift. These data are consistent with the reviewer’s second scenario (above) where non-H3K27me upregulated genes (and presumably indirect targets) show increased transcription without the +1 nucleosome shift.

1. Line 122: What is an "early condition phenotype"?

We’ve modified the sentence to clarify, as follows.

“We noticed that disruption of these two genes resulted in an early conidiation (production of asexual spores) phenotype that appeared as more dense growth in the spot tests (Figure 1C,D) but this was not accompanied by an increased linear growth rate.”

2. Lines 233-238: What is the conclusion drawn from this data?

We removed the original Figure 4F,G showing the position of H3K27me2/3 gains and losses in ∆iswi and ∆acf1 relative to the telomere as we felt these observations were tangential. The new Figure 3 now shows more relevant information regarding changes in H3K27me2/3 and changes in gene expression.

3. Figure 2: It is interesting that transcriptional de-repression is much more robust in iswi mutant than acf1. Given that ISWI purification did not yield other form of ISWI complexes, this is not because of other forms of ISWI-containing remodelers (CHRAC also depends on ACF1). So, another possibility is that, in an iswi mutant, ACF1 functions in a dominant negative fashion to derepress genes further. If this is the case, iswi acf1 double mutants should behave like an acf1 single mutant. The authors should test this for a few targets genes to establish how ACF functions.

While we are unable to make firm conclusions about ISWI-containing complexes with our mass spec data, we do feel they are consistent with ISWI forming multiple protein complexes, as it does in other organisms (see reviewer 1, comment 4). Therefore the most parsimonious explanation for the more robust transcriptional derepression in ∆iswi is because it is a member of multiple protein complexes. However, we do agree that it is interesting that genes that are upregulated in both iswi and acf1 are much more highly expressed when iswi is deleted. We tested the reviewer’s hypothesis by doing RT-qPCR on three genes that were more robustly upregulated in ∆iswi than ∆acf1 based on mRNAseq (NCU07152, NCU08796, NCU04860) in WT, ∆iswi, ∆acf1 and ∆iswi;∆acf1 double mutant strains. We found that gene expression in the double mutant was robust, similar to the ∆iswi single mutant strain, thus ruling out the dominant negative hypothesis. Because this yielded a negative result we did not include it in the manuscript.

4. Figure 2S1: ~1/3 of genes repressed by Set7 are not repressed by ACF. Do these genes have any special features?

We performed a transcription factor binding site motif analysis (+1 nucleosome +/- 50 bp) using HOMER in the subset of genes that are uniquely upregulated in ∆set-7 (n=23), but the program did not return any motifs predicted to be true positive binding sites.

5. Do ACF signals by Dam-ID mapping show something specific at H3K27me genes that are derepressed and/or show changes in nucleosome positioning? If the authors can show specific localization at these genes, it would help support the model that the transcriptional effects mediated by ACF are direct.

While we are able to detect a significant decrease in the ACF1 DamID signal in ∆set-7 at H3K27 methylated regions compared to non-H3K27 methylated regions of the genome, we are unable to detect a further significant decrease at the subset of H3K27 methylated genes that increase expression and have a nucleosome shift in ∆acf1.

6. ISWI regulates only a small subset of H3K27me genes (98 of 836). Are there unique properties of these genes that may contribute to our understanding of their mechanism of derepression? For example, are they regulated by a common transcription factor? Are their nucleosome positions in WT different from H3K27me genes not regulated by ISWI?

As illustrated in Figure 2 H, iswi regulates 180 (98+82) of H3K27-methylated genes. This is even more than the number that are derepressed by loss of the methyl mark itself (the set-7 mutant causes de-repression of 63 genes) and it is important to consider that some H3K27me-marked predicted genes may be “pseudogenes” or else regulated by unknown processes. We attempted to identify a transcription factor that worked with ACF by looking for a binding motif surrounding the +1 nucleosome (+1 nucleosome +/- 50 bp) in the subset of genes that were upregulated and showed a nucleosome shift in ∆iswi (n=93) and ∆acf1 (n=61). No motif was found to be enriched in these regions. By comparing the WT nucleosome positions in genes that are H3K27 methylated and upregulated in ∆acf1 or ∆iswi (Figure 6I, K) to those that are H3K27 methylated and not upregulated in ∆acf1 and ∆iswi (Figure 6J, L) we do notice some differences. We find that overall nucleosomes tend to be more well positioned (smoother, sharper peaks and valleys) in the genes that are not upregulated in these mutants. We also see a less pronounced NDR in the genes that are not upregulated. While these are interesting observations, it is hard to draw any biological conclusions from these differences.

[Editors’ note: what follows is the authors’ response to the second round of review.]

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

In this re-submitted manuscript, the authors addressed issues raised by the reviewers. As a result, this manuscript was substantially strengthened. The main conclusions from Figures1 through 3 are consistent with those of Kamei et al. (2021). However, this manuscript provides strong evidence that ACF directly functions to promote Polycomb-dependent repression through chromatin remodeling, which has not been demonstrated. In addition, PRC2/H3K27me-dependent ACF targeting is novel. Finally, the authors' model that facultative chromatin can be classified into several groups based on their dependence on SET7, ASH1, and ACF (Figure 6) is potentially important for guiding future research directions of the field.

The data that the effect of ACF on the position of +1 nucleosomes is distinct between H3K27me positive vs negative loci (Figure 5, Figure 5 Supplemental 2) is quite interesting. Is the degree of targeting of ACF and its effect on transcription similar between H3K27me positive and negative genes? If so, this result suggests currently unknown effects of H3K27me on ACF activity. In this case, Figure 5 Supplemental 2 should be move to the main Figure and the result deserves more extensive discussions as this would increase the impact of the manuscript. On the other hand, if the degree of ACF targeting and/or the level of transcription can explain the difference in the +1 nucleosome position, it needs to be stated in the text.

We performed further analyses to address the question of whether ACF targeting or level of transcriptional upregulation could explain the difference in ACF activity at genes that are upregulated in ∆acf1 strains with and without H3K27 methylation. We examined the ACF1-DamID signal at the +1 nucleosome (+/- 500bp). Upstream of the +1 nucleosome we found that the level of targeting was similar at genes that were methylated and unmethylated with a slight shift in the peak methylation, while downstream we saw ACF1-Dam targeting was slightly higher in H3K27-methylated regions. These findings were consistent with what we report for all genes (H3K27 methylated vs. non-methylated) in Figure 4. The level of transcription (log2FC ∆acf1/WT) was slightly higher (p=0.041) at H3K27-methylated genes compared to non-methylated genes. We did not find these data convincing enough to warrant inclusion in the manuscript. However, since the difference in ACF activity cannot clearly be ascribed to differences in targeting or transcriptional activity, following your suggestion, we decided to better highlight the data suggesting some specificity for H3K27-methylated regions. Thus, we moved the nucleosome plots showing no shift in non- H3K27 methylated that are upregulated in ∆iswi and ∆acf1 to the main Figure 5. To make room for this, we relegated two of the three analyses of the accessory factors (iaf-1 and iaf-2) to the supplementary figure (Figure 5 Sup2 panels C&D). We also made a minor change to the conclusion drawn from these data in the Results section and added a sentence regarding this in the Discussion.

https://doi.org/10.7554/eLife.77595.sa2

Article and author information

Author details

  1. Elizabeth T Wiles

    Institute of Molecular Biology, University of Oregon, Eugene, United States
    Present address
    Department of Molecular Biology and Biochemistry, University of California, Irvine, United States
    Contribution
    Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Supervision, Validation, Writing – original draft
    For correspondence
    tish.wiles@gmail.com
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-7269-7466
  2. Colleen C Mumford

    Institute of Molecular Biology, University of Oregon, Eugene, United States
    Contribution
    Data curation, Investigation, Validation, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-0213-0901
  3. Kevin J McNaught

    Institute of Molecular Biology, University of Oregon, Eugene, United States
    Present address
    Genapsys, Inc, Westminster, United States
    Contribution
    Investigation, Resources, Validation, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-6887-3161
  4. Hideki Tanizawa

    Institute of Molecular Biology, University of Oregon, Eugene, United States
    Contribution
    Conceptualization, Data curation, Formal analysis, Methodology, Software, Validation, Writing – review and editing
    Competing interests
    Kevin J. McNaught is affiliated with Genapsys, Inc The author has no financial interests to declare
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-2573-2473
  5. Eric U Selker

    Institute of Molecular Biology, University of Oregon, Eugene, United States
    Contribution
    Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing – review and editing
    For correspondence
    selker@uoregon.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-6465-0094

Funding

National Institutes of Health (GM127142)

  • Eric U Selker

National Institutes of Health (GM093061)

  • Eric U Selker

American Heart Association (14POST20450071)

  • Elizabeth T Wiles

National Institutes of Health (HD007348)

  • Kevin J McNaught

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

The authors thank J Lyle and R Morse for help in genetic mapping of UV-generated mutants; V Bicocca for performing initial experiments characterizing ISWI and for comments on the manuscript; and J McKnight and L McKnight for providing guidance and reagents for rapid MNase digestion. The authors also thank T Bailey, D Donovan, L McKnight, and K Noma for helpful comments on the manuscript. This work was funded by the National Institute of General Medical Sciences (GM127142 and GM093061 to EUS), American Heart Association (14POST20450071 to ETW), and KJM was partially supported by the National Institutes of Health (HD007348).

Senior Editor

  1. Kevin Struhl, Harvard Medical School, United States

Reviewing Editor

  1. Jerry L Workman, Stowers Institute for Medical Research, United States

Reviewer

  1. Jerry L Workman, Stowers Institute for Medical Research, United States

Publication history

  1. Preprint posted: June 30, 2021 (view preprint)
  2. Received: February 4, 2022
  3. Accepted: March 3, 2022
  4. Accepted Manuscript published: March 8, 2022 (version 1)
  5. Accepted Manuscript updated: March 9, 2022 (version 2)
  6. Version of Record published: April 25, 2022 (version 3)
  7. Version of Record updated: April 28, 2022 (version 4)

Copyright

© 2022, Wiles et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 991
    Page views
  • 148
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Chromosomes and Gene Expression
    2. Genetics and Genomics
    Mary Couvillion et al.
    Tools and Resources

    Understanding the complex network that regulates transcription elongation requires the quantitative analysis of RNA polymerase II (Pol II) activity in a wide variety of regulatory environments. We performed native elongating transcript sequencing (NET-seq) in 41 strains of S. cerevisiae lacking known elongation regulators, including RNA processing factors, transcription elongation factors, chromatin modifiers, and remodelers. We found that the opposing effects of these factors balance transcription elongation and antisense transcription. Different sets of factors tightly regulate Pol II progression across gene bodies so that Pol II density peaks at key points of RNA processing. These regulators control where Pol II pauses with each obscuring large numbers of potential pause sites that are primarily determined by DNA sequence and shape. Antisense transcription varies highly across the regulatory landscapes analyzed, but antisense transcription in itself does not affect sense transcription at the same locus. Our findings collectively show that a diverse array of factors regulate transcription elongation by precisely balancing Pol II activity.

    1. Chromosomes and Gene Expression
    2. Neuroscience
    Lesly Calderon et al.
    Research Article Updated

    Cohesin and CTCF are major drivers of 3D genome organization, but their role in neurons is still emerging. Here, we show a prominent role for cohesin in the expression of genes that facilitate neuronal maturation and homeostasis. Unexpectedly, we observed two major classes of activity-regulated genes with distinct reliance on cohesin in mouse primary cortical neurons. Immediate early genes (IEGs) remained fully inducible by KCl and BDNF, and short-range enhancer-promoter contacts at the IEGs Fos formed robustly in the absence of cohesin. In contrast, cohesin was required for full expression of a subset of secondary response genes characterized by long-range chromatin contacts. Cohesin-dependence of constitutive neuronal genes with key functions in synaptic transmission and neurotransmitter signaling also scaled with chromatin loop length. Our data demonstrate that key genes required for the maturation and activation of primary cortical neurons depend on cohesin for their full expression, and that the degree to which these genes rely on cohesin scales with the genomic distance traversed by their chromatin contacts.