H3K4me2 ChIP-Seq reveals the epigenetic landscape during mushroom formation and novel developmental regulators of Schizophyllum commune

Mushroom formation represents the most complex multicellular development in fungi. In the model mushroom Schizophyllum commune, comparative genomics and transcriptomics have previously resulted in a regulatory model of mushroom development. However, little is known about the role of epigenetic regulation. We used chromatin immunoprecipitation sequencing (ChIP-Seq) to determine the distribution of dimethylation of lysine 4 on histone H3 (H3K4me2), a mark for transcriptionally active genes, during monokaryotic and dikaryotic development. We identified a total of 6032 and 5889 sites during monokaryotic and dikaryotic development, respectively. The sites were strongly enriched near translation initiation sites of genes. Although the overall epigenetic landscape was similar between both conditions, we identified 837 sites of differential enrichment during monokaryotic or dikaryotic development, associated with 965 genes. Six transcription factor genes were enriched in H3K4me2 during dikaryotic development, indicating that these are epigenetically regulated during development. Deletion of two of these genes (fst1 and zfc7) resulted in arrested development of fruiting bodies, resulting in immature mushrooms. Together these results indicate that H3K4me2 ChIP-Seq is a powerful new tool to map the restructuring of the epigenetic landscape during mushroom development. Moreover, it can be used to identify novel developmental regulators.

minutes. At this point, take the previously stored input control samples and adjust the volume to 500 µL with water. To remove the RNA, add 50 µg RNAse A to each sample and incubate at 50 °C for 1 hour. Then proceed with de-crosslinking by adding 75 µL reverse crosslinking buffer (1.25 M NaCl, 250 mM Tris-HCl pH 6.5, 62.5 mM EDTA, 5 mg mL -1 proteinase K) and incubate at 65 °C overnight.

DNA isolation
To each sample add 1 volume of phenol-chloroform (1:1), mix thoroughly, and centrifuge the samples at 15,000 g for 5 minutes. Transfer the upper aqueous phase to a new tube. Repeat this step until no white interphase is present. Note: depending on the amount of input, this step must be repeated up to 5 times. Especially in the input control samples a lot of protein is present at the start of DNA isolation.
To remove phenol, add 1 volume of chloroform, mix thoroughly, and centrifuge the samples at 15,000 g for 5 minutes. Transfer the aqueous phase to a new tube and proceed with DNA precipitation. To each sample add 0.1 volume of 3M NaAC pH 5.6 and 2 volumes of ethanol, mix thoroughly, and incubate the samples at -80°C for 2 hours. Note: 20 µg of glycogen can be added during this step to improve DNA precipitation and make the pellet easier to observe.
Centrifuge the samples at 4°C for 45 minutes at 15,000g and discard the supernatant carefully. Resuspend the precipitated DNA in 1 mL 70% ethanol and mix thoroughly. Collect the DNA by centrifugation at 15,000 g for 15 minutes at 4°C and discard the supernatant. Briefly dry the pellet and resuspend in 30 µL TE. Clean up the DNA further by using the isolated DNA as input in the purification method of choice. Note: We use the ChargeSwitch gDNA plant kit from ThermoFisher Scientific for further cleanup, as S. commune produces a lot of polysaccharides that contaminate the gDNA and are not eliminated by traditional purification methods, including phenol-chloroform extraction and silica-based methods. Early ChIP-qPCR trials without additional purification did not yield reproducible results, which was alleviated by DNA cleanup.

ChIP-qPCR
To determine if the ChIP DNA is enriched compared to the input control qPCR can performed at regions expected to be enriched with the modification of choice and regions that are expected not to be enriched. Compare the ΔCt of the positive and negative regions in both input control samples and the ChIP samples. The difference in ΔCt (ΔΔCt) between the input control and ChIP samples is a measure for enrichment. Note: We used qPCR on the region near the translation initiation site of actin, β-tubulin and gpd as regions with expected enrichment. Negative regions were selected that showed no expression nearby in any RNA-Seq data of S. commune. Specifically: scaffold 10: 864,508-864,623 (region A); scaffold 12: 8,012-8,112 (region B); scaffold 19: 17,635-17,749 (region C). For each region we found a 30-37-fold enrichment of active regions compared to regions with no expression (Supplemental figure S8).

Library preparation and sequencing
The libraries were generated with the NEXTflex Rapid DNA-Seq Kit Bundle (Bioo Scientific, TX, USA) according to manufacturer's specifications. The resulting libraries were sequenced on the Illumina NextSeq500 2x75 mid output platform. During library preparation, the ChIP DNA can be amplified by PCR to obtain a suitable amount for sequencing. However, this may introduce bias and impact the significance of data. It is therefore recommended to increase the amount of input material when possible. Furthermore, it is strongly recommended to use paired-end sequencing, as the algorithm for peak prediction can use this information to determine the average fragment-length. This improves peak detection and results in more reproducible data.

Sequencing analysis
Reads can be trimmed using the standard tools and aligned with Bowtie2 3 . We used the sensitive-local option for alignment. Next, filter the reads based on paired-end alignment and a quality score >1 with samtools 4 . This is a very low cut-off and results in only a small loss of reads. To remove any bias introduced by PCR amplification, it is recommended to mark and remove any duplicate reads from the alignments. A strong tool for this is Picard MarkDuplicates 5,6 , which can identify optical duplicate reads and marks each duplicate with a 0x400 flag that can be filtered out with samtools. Note: for our samples only a small number of reads was excluded. However, for less ubiquitous modifications, more amplification may be required to obtain sufficient DNA for sequencing, resulting in more optical duplicate reads.  Figure     However, between controls and ChIP samples there is very low correlation (< 0.10). Replicates of the ChIP samples had a higher correlation than between the samples.  Table 4. ChkA and D primers bind outside the gene deletion in both the wildtype and deletion strains, while ChkB and C are WT-specific. Nour_ChkB and nour ChkC are deletionmutant specific (A). PCR should yield a band of ~1 kb in the wildtype using the WT primers, and in the gene deletion using the KO primers. The KO primers of the right flank of zfc7 detect a non-