A crucial role for dynamic expression of components encoding the negative arm of the circadian clock

In addition to its role as the major transcription factor for frq, WC-1 is also the principal blue-light photoreceptor for the organism, directly or indirectly mediating light-induced gene expression⁴⁷. Interestingly, the ∆brd-8 strain that displayed a ~25-h period and delayed dark FRQ expression retains its capacity to detect light to induce expression of light-responsive genes (Supplementary Figure 3). Unexpectedly, light-induced expression of some genes (e.g. frq, vvd, csp-1, and sub-1) is even significantly higher in ∆brd-8 than in WT, consistent with differing roles for BRD-8 depending on the promoter element (C-box vs. pLRE) used.

To understand the role of BRD-8 at a mechanistic level, core circadian component expression was examined by Western blotting (Fig. 1c). FRH and WC-2 levels are normal or slightly reduced in ∆brd-8, and consistent with the long period of ∆brd-8, new FRQ in ∆brd-8 appears later than that in WT and is slightly reduced. However, the level of WC-1 in ∆brd-8 is significantly reduced. Active WC-1 is unstable and FRQ, by inactivating WCC, stabilizes WC-1^5,6, so to test whether the low WC-1 level is caused by a defective negative arm, ∆brd-8 was backcrossed to the loss of function allele frq⁹; WC-1 levels are further reduced in the frq⁹; ∆brd-8 double mutant (Fig. 1d), which indicates that brd-8 regulates WC-1 expression at least partially independently of the negative feedback loop. Reduced WC-1 levels as seen in ∆brd-8 can result in modestly long circadian periods⁴⁸. To test whether the long circadian period in ∆brd-8 is caused by the reduced WC-1 level, the native promoter of wc-1 in ∆brd-8 was replaced by the strongly inducible qa-2 promoter. In the presence of inducer (quinic acid, QA), the wc-1 and WC-1 levels are raised to those in WT (Fig. 1e, f) or even higher, but the period defect of ∆brd-8 is not rescued (Fig. 1g), indicating that the longer period in the mutant is not caused by the low WC-1 level. Similarly, to test whether changes in FRQ levels underlie the long period of ∆brd-8, a second copy of the frq gene with its native promoter was inserted at the csr locus of ∆brd-8. The long period of ∆brd-8 was rescued to WT (Fig. 1h), indicating that reduced frq expression in ∆brd-8 is the cause of the long period length.

If loss of BRD-8 impacts NuA4 function leading to period lengthening, then loss of other NuA4 subunits or associated proteins might be expected to have similar and additive effects. This was tested; ∆eaf-3 shows an impaired long-period core clock by a luciferase assay (Fig. 2d) and a strong growth defect by race tube analysis (Fig. 2e). ∆bye-1 similarly displays a circadian period lengthened by ~2 hrs (Fig. 2e), while disruption of both eaf-3 and bye-1 leads to low and arrhythmic luciferase signals (Fig. 2d). The ∆brd-8, ∆bye-1 strain bearing deletions in both components shows a period lengthened to the same degree as the single mutants consistent with them working through the same pathway (Fig. 2d). Interestingly, in the perithecia, transcripts of all NuA4 subunits as well as BYE-1 are subject to multiple A-to-I editing events (Fig. 2b)⁵¹. This is of potential interest because, whereas most editing sites in animals occur in noncoding regions associated with repetitive elements, editing in fungi is typically non-synonymous, is under positive selection, and can often lead to changes in coding or STOP sites⁵².

To identify motifs in BRD-8 essential for its circadian function and intra-complex interactions, strains bearing a series of deletions of brd-8 (Fig. 5d) were generated and analyzed by race tube and Western blotting assays. Western blotting analyses showed that of the four brd-8 deletion strains, only ∆1030-1306 has a reduced protein level (Supplementary Figure 4d). Two of these mutants, missing aa 2–348 and aa 703-1029, showed a conidiation phenotype similar to ∆brd-8 (Fig. 5e), suggesting that these two regions are required for its circadian function. Although analysis of BRD-8’s open reading frame (ORF) revealed a region of bromodomain near the C-terminus of the protein (Fig. 5d), deletion of aa 1030–1306 within which the bromodomain is located resulted only in a slightly longer period (23.4 hrs) (Fig. 5e, f), which suggested that the domain is not essential for its circadian function. This result is consistent with the existence of multiple BRD-8-containing complexes in the cell, only some of which require the bromodomain, and that those requiring the bromodomain are not the functions relevant to circadian timekeeping.

Because ESA-1 and BYE-1 were identified from the BRD-8 interactome analyses and disruption of either gene results in a lengthened circadian period, we checked by immunoprecipitation whether the interaction between BRD-8 and ESA-1 or BYE-1 is influenced in the brd-8 mutants. Interestingly, BRD-8^∆2-348, in which a proline-rich low complexity region is deleted, binds normally to BYE-1 but does not interact with ESA-1 (Fig. 5g), while BRD-8^∆703-1029 associates with ESA-1 but fails to complex with BYE-1 (Fig. 5h), consistent with BRD-8 being the adapter protein for ESA-1 and BYE-1. brd-8^∆349-702 and brd-8^∆1030-1306 showed normal interactions with ESA-1 and BYE-1 (Fig. 5g, h) and displayed ~WT circadian phenotypes (Fig. 5e, f).

Because the circadian clock in Neurospora globally regulates gene expression, we investigated whether the expression of BRD-8, BYE-1, and several additional NuA4 subunits are also under circadian control. Promoters of brd-8, bye-1, eaf-3, vid-21, and esa-1, along with histoneh2a.z, and histoneh4-1 were fused with the codon-optimized LUC gene respectively and transformed to the WT strain at the csr locus. The promoters of brd-8, bye-1, eaf-3, vid-21, and histoneh2a.z genes were shown to drive rhythms in LUC expression (Fig. 6f), and ARP-4 was found to be rhythmic in a separate study⁵⁵, while no obvious circadian rhythmicity was seen for the esa-1 or histoneh4-1 promoter (Supplementary Fig. 8b). Furthermore, to better understand and analyze brd-8 dynamics in vivo, we constructed a brd-8-luciferase translational fusion reporter that directly tracks BRD-8 protein expression and transformed it into the WT strain. The protein level of BRD-8-LUC oscillated in vivo in a circadian manner (Fig. 6f), confirming that brd-8 is a ccg (clock-controlled gene), and the strain bearing BRD-8-LUC showed an overt rhythm with a WT period (Fig. 6g), indicating that LUC does not impact its function. The data are consistent with a model in which an BRD-8/BYE-1/NuA4 complex controls frq expression and thereby the clock, perhaps by modulating transcription elongation, while several subunits of the BRD-8/BYE-1/NuA4 complex are also themselves under circadian control. In this way, elements of the BRD-8/BYE-1/NuA4 complex both regulate the clock and are regulated by the clock, contributing to a feedback loop surrounding the core circadian oscillator.

In this study several subunits associating with the NuA4 complex were identified including BRD-8 and BYE-1 (Fig. 7c). In yeast, the NuA4 complex is comprised of 13 subunits (Fig. 2c). In Neurospora, however, no gene having homology to eaf-5 exists, and indeed orthologs of EAF5 are found only in Saccharomyces and closely related species (Supplementary Fig. 5). Among the aspects of physiology possessed by fungi outside of the Saccharomycotina are responses to environmental stimuli including light and time-of-day, circadian biology. This correlation suggests the possibility that multiple NuA4 complexes with more subunits may have evolved to support additional aspects of physiology.

Another aspect of regulation found in Neurospora but not Saccharomyces is RNA A-to-I editing. In the fungi this mainly occurs in coding regions and results in nonsynonymous changes to proteins encoded⁵¹, which dramatically extends the proteomic complexity during translation and thus can broadly impact cellular and physiological events. Interestingly, mRNAs of all NuA4 subunits undergo multiple A-to-I editing events in the sexual phase (Fig. 2b), which may enable histone H4 acetylation to epigenetically control the chromatin state favoring sexual development and adaptation by influencing chromatin events such as gene expression, DNA repair, and genome duplication.

BRD-8 influences rhythmicity through chromatin events. In ∆brd-8, nucleosome density is higher at frq while histone H4 is less acetylated than in WT, both of which effects will impede frq transcription. frq is known to lie in a region characterized by very rapid nucleosome turnover¹¹ so in affecting nucleosome density and modification, loss of brd-8 might be expected to cause reduced frq expression and a long period which can be rescued by adding a second copy of the frq gene to the ∆brd-8 strain. The data are consistent with models in which timely expression of frq is required for bringing kinases to phosphorylate WCC and effect repression in the circadian cycle. As FRQ and WC-1 are present in similar amounts and FRQ-promoted WCC phosphorylation is a slow process, at least in part because FRQ is predominantly cytoplasmic⁹, the level of FRQ needs to be sufficient to bring about repression to WCC.

BYE1 was originally identified in Saccharomyces through a mutation that bypassed the requirement for the essential prolyl isomerase ESS1²⁴. Ess1-induced conformational changes are seen as attenuating Pol II elongation, thus helping to coordinate the assembly of protein complexes on the Pol II CTD and thereby regulating transitions between multiple steps of transcription²⁴. BYE1 binds directly to Pol II during early transcription elongation and tethers surrounding histones containing active post-translational modifications (PTMs), perhaps to prevent loss of histones during polymerase passage through chromatin²⁶ but also slowing transcription. Although in Saccharomyces BYE1 works as a free protein, its action(s) in Neurospora in conjunction with NuA4 may be similar to those seen in yeast, acting on elongation through regulating Pol II CTD or just enhancing ESA-1 activity on chromatin, an activity not reported in Saccharomyces. In either case, a seeming paradox in the circadian data related to loss of BRD-8 and BYE-1 arises from the facts that (1) loss of BRD-8 which results in increased nucleosome density is expected to slow transcription but (2) loss of BYE-1, a negative regulator of transcription elongation²⁶, would be expected to increase transcription elongation; yet both result in period lengthening. One clue to this lies in the data from Fig. 5 showing that BRD-8 forms the bridge between ESA-1 and BYE-1; hence, loss of BRD-8 results in parallel loss of BYE-1 from the NuA4 complex. This still leaves unexplained, however, how single loss of BYE-1 in the presence of normal BRD-8 results in modest period lengthening. A possible resolution to this lies in the data of Fig. 6 which detail the dosage effects of BRD-8 on rhythmicity. While complete loss of BRD-8 slows period to the greatest extent, the highest expression of BRD-8 also results in modest period lengthening, comparable to loss of BYE-1. It may be that overexpression BRD-8 has consequences on transcription similar to loss of BYE-1, and that normal rhythmicity requires a happy medium of regulators of transcription elongation, allowing it to proceed but not to proceed so quickly as to interfere with ordered assembly of protein complexes on the Pol II CTD.

The unexpected identification of a human BRD8-like protein associated with a fungal NuA4 acetyltransferase complex, and the role of acetylation in the fungal clock, warrant comment. In mammals, Brd8 is expressed in multiple isoforms, some with one and some with two bromodomains, although expression of the two-bromodomain isoform is restricted to testes, whereas the single bromodomain isoform is ubiquitously expressed (e.g., GTExportal.org) and it is similar in size to Neurospora BRD-8 that likewise has a single bromodomain. While as noted above, BRD8 is associated with multiple mammalian complexes including NuA4/TIP60, the association of a bromodomain-containing protein in a fungal NuA4 complex has been debated (reviewed in refs. ^49,71). Bdf1 in the Saccharomyces cerevisiae SWR1 complex, has a single bromodomain as does Bdc1, a component of the NuA4 histone acetyltransferase complex In Saccharomyces pombe (reviewed in refs. ^49,71). However, the overall amino acid identity of Neurospora BRD-8 with Bdf1 and Bdc1 is relatively low (6.96% and 3.9%, respectively, similar to background), Bdf1 and Bdc1 are a third of the size of BRD-8, and perhaps most significantly, neither Bdf1 nor Bdc1 are BLASTP hits with BRD-8. Bdf1 is, however, a component of the yeast SWR complex which contains additional components orthologous to mammalian NuA4/TIP60 components, prompting the observation that the human NuA4 complex corresponds to a near-perfect fusion of the two distinct complexes in yeast, namely NuA4 and SWR1 (Doyon and Cote, 2004). This prompted us to re-examine the mass spectrometry data (Fig. 2 and Supplementary Data 1) for presence of putative SWR complex components. No peptides from subunits exclusively in SWR (i.e. orthologs of Bdf1, Vps72, Rvb1, Rvb2) were recovered from the BRD-8^V5 purification (Supplementary Data 1), suggesting that the presence of BRD-8 in association with NuA4 is more like that seen in mammals than the association of bromodomain-containing proteins sometimes seen in yeasts.

A second point of interest sparked by the identification of BRD-8 in association with NuA4 is the potential role of acetylation in rhythmicity. This question is nicely set up by the elegant study of Petkau et al. who showed in mammals that the NuA4/TIP60 complex acetylates BMAL1 (the ortholog of Neurospora WC-1), providing a docking site that brings the BRD4-P-TEFb complex to DNA-bound BMAL1, promoting release of Pol II and elongation of circadian transcripts and revealing a role for control of transcriptional elongation in the mammalian clock²³. At first glimpse there are clear parallels to the BRD-8 story here, in particular because WC-1 has been reported to be acetylated⁷²; however, the similarities weaken when examined more closely. Neurospora does have a Brd4 ortholog (BDP-3 encoded by NCU08423) and, unlike BRD-8, it is essential as is Brd4 in humans. The acetylation of WC-1 appears only to be important for light-regulation⁷², and perhaps most importantly, the structure/function data from Fig. 5d–h plainly indicate that the bromodomain of BRD-8 is not required for its circadian function. The mass spectrometry data from Supplementary Data 1 suggest that BRD-8 may interact with a number of different proteins and complexes, and if so it is like human Brd8 which interacts with at least 135 proteins (https://thebiogrid.org/116108↗), is ubiquitously expressed, and has roles in a host of processes; in both systems it seems plausible that not all of these roles would require the bromodomain. In the present case, BRD-8 uses its unstructured region to interact with BYE-1, and deleting its single bromodomain still results in a normal period, while removal of two disordered domains individually phenocopies the whole gene knockout of brd-8 (Fig. 5d–h). Although BRD-8 provides no evidence for a role of bromodomain-mediated acetylated residue binding, a more remote route to achieve this must be mentioned. The human ortholog of BYE-1 is DIDO1 (death inducer-obliterator 1) which has been shown in a high throughput affinity capture-mass spectrometry (ms) screen⁷³ to interact with Brd4, the protein shown by Petkau et al.²³ to mediate the pause/release of Pol II from DNA-bound BMAL1. If the same mechanism worked in Neurospora, it should be mediated by the Neurospora ortholog of Brd4 (BDP-3) and be independent of the BRD-8/BYE-1 association, and if so the period lengthening seen from loss of either BRD-8 or BYE-1 alone should be additive in the double mutant ∆brd-8, ∆bye-1; however, this is not the case (Figs. 1a, and 2d). Additionally, there is no evidence (mass spectrometry in Supplementary Data 1 or targeted immunoprecipitation of BRD-8 followed by blotting for FRQ or WCC) for interaction of BRD-8 with FRQ or WCC. Lastly BDP-3 (the Neurospora ortholog of human BRD4) does not appear in the mass-spectrometry interactome data (Supplementary Data 1), indicating that even if BYE-1 has a circadian role through BDP-3, it cannot be linked in any mechanistic way to BRD-8.

A wealth of research in recent years has established mechanisms through which epigenetic regulation plays an important role in controlling circadian systems (reviewed in refs. ^74–78); however, the reverse has rarely been confirmed (reviewed in⁷⁹). Here we have shown that expression of histoneh2a.z, as well as a number of NuA4 subunits including brd-8 gene and protein are circadianly controlled (Fig. 7c), linking the circadian clock to the basic cellular machinery. This prompted us to examine online databases for circadianly regulated genes in mammals (CIRCA: Circadian gene expression profiles; http://circadb.hogeneschlab.org↗) for evidence of circadian regulation of mammalian NuA4/TIP60 components. Interestingly, strong albeit sometimes low amplitude rhythms were seen for gene expression of brd8 as well as Trrap, RUVBL1, and DMAP1. This suggests that in mammals, as in Neurospora, the events surrounding histone acetylation and transcription elongation describe a feedback loop surrounding the core clock that serves to amplify and stabilize the rhythm, contributing to persistence. These data also provide evidence for a mechanism through which the circadian clock can control gene expression through an indirect mechanism not directly mediated by the WCC or CLOCK/BMAL1 complex, the principal means of circadian output control.

328-4 (ras-1^bd, A) was used as a clock-WT strain in the race tube analyses and 661-4a (ras-1^bd, A, his-3::C-box-driven luc), WT in the luciferase assay; this contains the frq C-box fused to codon-optimized firefly luciferase (transcriptional fusion) at the his-3 locus. Neurospora transformation was performed as previously reported⁶². To test the rhythmicity of gene expression, promoter-driven luciferase (LUC) reporter constructs were generated by transforming four PCR products [5′ of csr, promoter region of gene, codon optimized luciferase sequence, and 3′ of csr] with digested plasmid (pRS426) into yeast⁸⁰. Race tube medium contains 1 × Vogel’s salts, 0.17% arginine, 1.5% bacto-agar, and 50 ng/mL biotin with glucose at 0.1%, and liquid culture medium (LCM) contains 1 × Vogel’s, 0.5% arginine, 50 ng/mL biotin, and 2% glucose. Quinic acid (QA) was added into race tube medium with certain concentrations as indicated. Unless otherwise specified, race tubes were cultured in constant light for 16–24 h at 25 °C and then transferred to the dark at 25 °C⁸¹. BRD-8, BYE-1, and NuA4 deletion strains generated by the Neurospora genome project were obtained from Fungal Genetics Stock Center (FGSC).

Procedures for preparation of protein lysates and Western blots (WB) were followed as described in refs. ^8,82 For WB, 15 µgs of whole-cell protein lysate were loaded per lane in a 3-8% tris-acetate or 4-12% bis-tris SDS gel. Custom rabbit antibodies used in this study include antisera against WC-1 (custom antibody ordered from Pocono Rabbit Farm and Laboratory; depleted and used in Western blotting at a dilution of 1:250)⁸³, WC-2 (custom antibody ordered from Pocono Rabbit Farm and Laboratory; used at a dilution of 1:5,000)⁸⁴, FRQ (custom antibody ordered from Pel-Freez Biologicals; depleted and used at a dilution of 1:250)¹⁰, and FRH (custom antibody ordered from Pocono Rabbit Farm and Laboratory; used at a dilution of 1:5,000)⁵. V5 antibody (Thermo Pierce), FLAG antibody (Sigma-Aldrich), and HA (Abcam, ab9110) were diluted at 1:5,000 for use as the primary antibody. All uncropped and unprocessed whole blots used for figure preparations were deposited into the Source Data file except for the ones for Supplementary Figure 4a, which were damaged permanently in a lab fire.

VHF-tagged BRD-8 were purified with the same method applied for isolation of C-terminal VHF-tagged WC-1^8,62. 30 g of Neurospora tissue (brd-8^VHF) was used for purification of BRD-8. To exclude the possibility that associations with chromatin impacted identification of BRD-8-interactors in purifications, we (1) vortexed lysate at the top speed to mechanically disrupt chromatin: 10 s on/10 s off (put on ice) for a total of 2 min; (2) extensively sonicated protein lysate; (3) included benzonase nuclease in the protein lysis buffer; benzonase is widely used for the removal of nucleic acids (both DNAs and RNAs) from protein samples^85–88. To recover potential weak interactors, 5 g of tissue from brd-8^V5 were purified with V5 antibody (Thermo Pierce)-conjugated Dynabeads (Life Technologies). A single protein band or the whole lane⁸⁹ was cut off from the gel as indicated in the figure and used for mass spectrometry analysis.

Affinity purification of BRD-8’s interactors using VHF-tagged BRD-8 (the left gel in Fig. 2a) was performed twice (two biological replicates) with similar results, and identification of proteins from excised bands was carried out once with mass spectrometry. Interactor identification from BRD-8^V5 (the middle and right gels in Fig. 2a) was performed three times (three biological replicates), similar band profiles were observed, and specific bands or whole gel lanes were analyzed once by mass spectrometry. Validations for interaction of BRD-8 with its interactors identified from mass spectrometry analyses were performed with strains bearing epitope-tagged target proteins, and results were shown in Fig. 2c. One negative control used in the BRD-8’s interactor identification was an untagged Neurospora strain (328-4 [ras-1^bd, A]). Six samples were analyzed by mass spectrometry.

Samples were separated by SDS-PAGE gel electrophoresis and Coomassie-stained. Gel bands or lanes were excised, destained, and digested with trypsin in 50 mM ammonium bicarbonate overnight at 37˚C. Peptides were extracted using 5% formic acid / 50% acetonitrile (ACN), desalted, and dried. Peptides were analyzed on a Fusion Orbitrap mass spectrometer (Thermo Scientific) equipped with an Easy-nLC 1000 (Thermo Scientific).

Peptides were resuspended in 5% methanol / 1% formic acid and loaded on to an analytical resolving column (35 cm length, 100 μm inner diameter, ReproSil, C18 AQ 3 μm 120 Å pore) pulled in-house (Sutter P-2000, Sutter Instruments, San Francisco, CA) with a 45-min gradient of 5–34% LC-MS buffer B (LC-MS buffer A: 0.0625% formic acid, 3% ACN; LC-MS buffer B: 0.0625% formic acid, 95% ACN). The Fusion Orbitrap was set to perform an Orbitrap MS1 scan (R = 120 K; AGC target = 2.5e5) from 350 to 1,500 m/z, followed by HCD MS2 spectra detected by Orbitrap scanning (R = 15 K; AGC target = 0.5e5; max ion time = 50 ms) for 2 sec before repeating this cycle. Precursor ions were isolated for HCD by quadrupole isolation at width = 0.8 m/z and HCD fragmentation at 29 normalized collision energy (NCE). Charge state 2, 3, and 4 ions were selected for MS2. Precursor ions were added to a dynamic exclusion list ± 10 m ppm for 15 s.

Raw data were searched using COMET⁹⁰ (release version 2014.01) in high resolution mode against a target-decoy (reversed)⁹¹ version of the Neurospora proteome sequence database (UniProt; downloaded 3/2017) with a precursor mass tolerance of +/− 1 Da and a fragment ion mass tolerance of 0.02 Da, and requiring fully tryptic peptides (K, R; not preceding P) with up to three mis-cleavages. Static modifications included carbamidomethylcysteine and variable modifications included: oxidized methionine. Searches were filtered using orthogonal measures including mass measurement accuracy (+/− 3 ppm), Xcorr for charges from +2 through +4, and dCn targeting a < 1% FDR at the peptide level.

The list of identified interactors (Supplementary Data 1) was searched manually for proteins whose peptides were specifically enriched in the sample of BRD-8^V5 but not in the negative control (328-4, untagged). In the search for BRD-8^V5’s interactors, no statistical tests were taken, because we worried that this would rule out certain real targets due to their potential digestion or/and coverage issues in mass spectrometry analysis. Instead, we validated the identified interactors individually by immunoprecipitation assays with epitope-tagged proteins (Fig. 2c).

IP was performed as previously described^92,93. Briefly, 2 mgs of total protein were incubated with 20 μL of V5 agarose (Sigma-Aldrich, Catalog # 7345) or FLAG M2 resin (Sigma-Aldrich, Catalog # A2220) as indicated in the figures, rotating at 4 °C for 2 hrs. The agarose beads were then washed twice with the protein extraction buffer (50 mM HEPES [pH 7.4], 137 mM NaCl, 10% glycerol, 0.4% NP-40) and eluted with 50 µL of 5 × SDS sample buffer at 99 °C for 5 min.

ChIP experiments were done as previously described using fresh tissues⁶². Primer sets against C-box and pLRE were the same as in a previous publication⁸. Primer sets against the frq locus used in Fig. 4a were derived from a previous publication⁹⁴. “A” in the first “ATG” of the full-length frq coding region is designated as “+1” and the upstream nucleotides are labeled “-“ accordingly (without “0”). frq C-box is from −2,878 to −2,624 nt, the amplicon with the frq1 primer set is from −2,754 to −2,644 nt, while that by the C-box primer set is from −2,790 to −2,678 nt. ChIP using Neurospora tissue: Neurospora tissue was cross-linked with 3% formaldehyde for 15 min and quenched with 0.25 M glycine for 5 min. Tissue was then washed with PBS three times and vacuum-dried; 0.4 g was weighed, cut into 9 pieces, and each piece soaked in 0.5 ml SDS lysis buffer (50 mM Tris/HCL [pH 8.0], 1% SDS, 5 mM EDTA) containing Roche protease inhibitors (Sigma-Aldrich, Catalog # 11836170001). The soaked tissue was first sonicated for 8 sec for three times at 30% amplitude with a Bronson sonicator equipped with a microtip, and then further sonicated in a water-bath sonicator 5 times for 5 min each with an interval of 30 sec on and 30 sec off. The sonicated cell lysate was cleared of cellular debris by an 8,000 rpm centrifugation twice for 5 min each. 200 μl supernatant was saved and added to 1.8 ml RIPA buffer; 15 μL protein A magnetic beads and 2 μL V5 antibody (Abcam, Catalog # ab9116) were added to the mixture followed by rotation at 4 °C overnight. The following morning, immunoprecipitated DNA was washed with buffers A-D (buffer A: 20 mM Tris/HCL [pH 8.0], 0.1% SDS, 2 mM EDTA, 1% TritonX-100; buffer B: 20 mM Tris/HCL [pH 8.0], 0.1% SDS, 2 mM EDTA, 1% TritonX-100, 500 mM NaCl; buffer C: 10 mM Tris/HCL [pH 8.0], 0.25 M LiCl, 1 mM EDTA, 1% TritonX-100, 1% sodium deoxycholate; buffer D: 10 mM Tris/HCL [pH 8.0], 1 mM EDTA) once each, eluted with Elution buffer (0.1 M NaHCO₃, 1% SDS), reverse crosslinked with NaCl at the final concentration of 0.2 M, and purified with Qiagen PCR purification kit.

The purified ChIP DNA was end-repaired and ligated to Ion-compatible barcode adapters using Ion XpressTM Plus Fragment Kit (Catalog # 4471269) and Ion XpressTM Barcode Adapters (Catalog # 4471250). The final libraries were purified with two rounds of AMPure XP Bead capture to size select fragments between 160 to 340 bps in length. The emulsion clonal bead amplification to generate bead templates for the Ion Torrent platform was performed on the Ion Chef System (Thermo Fisher Scientific) with the Ion PI™ Hi-Q™ Chef Kit (Thermo Fisher Scientific, Catalog # A27198) and Ion PI™ Chip Kit v3 (Thermo Fisher Scientific, Catalog # A26771). Sequencing was done using the Ion PI™ Hi-Q™ Sequencing 200 Kit (Thermo Fisher Scientific, Catalog # A26772) on Ion Proton sequencer with sequencing data processing using the Torrent Suite TM Software (Ver. 4.0.2) on the Torrent server.

Samples were mapped to the Neurospora crassa genome (from GenBank) using TMAP (Ion Software, https://github.com/iontorrent/TMAP↗). Mapping efficiency and read quality were visualized using htseq-qa (http://htseq.readthedocs.io/en/release_0.9.1/↗). Peaks were called using MACS2 (https://github.com/taoliu/MACS↗), and peaks were annotated with ChIPseeker (Bioconductor, https://bioconductor.org/packages/release/bioc/html/ChIPseeker.html↗). In addition to MACS2, the SICER software (https://home.gwu.edu/~wpeng/Software.htm↗) was used to identify peaks and identify differentially abundant peaks between WT and mutant strains. For comparison of WT vs. control in histone and acetylated histone, three runs apiece were performed with window sizes of 200, 400 and 600, and gap sizes of 600, 1,200 and 1,800, respectively. Redundant peaks between the three runs (within 2,500 bp) were eliminated with a simple R script. Another script was used to identify genes within 1,000 bp from either border of the peaks.

The histone acetylation assay was performed as previously described with modifications⁹⁵. Briefly, ESA-1 was immunoprecipitated from the esa-1^V5 strain as described in the section of “IP” with V5 antibody-conjugated resins. Immunoprecipitated- ESA-1^V5-bound resins were washed twice with phosphate buffered saline containing 0.1% Tween 20 and once with the acetyl-transferase assay buffer (50 mM Tris-Cl [pH 8.0], 10% glycerol, 10 mM butyric acid, 0.1 mM EDTA, 1 mM DTT, 1 mM PMSF). 50 μL of the acetyl-transferase assay buffer containing 10 μM acetyl CoA (Sigma-Aldrich) was added to the washed resins. Recombinant human histone H2A (New England BioLabs, Catalog # M2502S) or H4 (New England BioLabs, Catalog # M2504S) respectively was added to the mixture at a final concentration of 0.1 mg/mL. Acetylation reactions were incubated at 30 °C for 1 hr, followed by adding 50 μL of 5 x SDS-PAGE sample buffer, heated at 99 °C for 5 min, electrophoresed in a 4-12% bis-tris SDS-PAGE gel, and transferred to a PVDF membrane. Acetylated histones H4 and H2A were detected by immunoblotting with rabbit anti-acH4 (UBI, Catalog # 06–866, 1:1000 dilution) and anti-acH2A K9 (Millipore Sigma, Catalog # 07-289), respectively.

Neurospora mRNA was isolated with TRIzol reagent (Life Technologies) and cDNA synthesis was done using SuperScript III first strand synthesis kit (Life Technologies) with 3 µg of purified RNA⁴⁷. Real-time PCR was performed with QuantiTect SYBR green RT-PCR kit (Qiagen) in an ABI 7500 Fast system. Primers against frq and wc-1 genes were from previous publications^47,96. Normalization for RT-qPCR was to NCU08964, a gene identified previously as among the most constantly expressed under different conditions and times or day, and least responsive to changes in media⁴⁰.

Mass spectrometry was performed as previously described⁶². Luciferase assays were performed as previously described^8,81. If QA was used in the luciferase assay, it was indicated in the figures. Nuclear preparation was performed as reported⁹².

Further information on research design is available in thelinked to this article. Nature Portfolio Reporting Summary

Supplementary Information Peer Review File Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 Reporting Summary