Analysis of the Chromosomal Localization of Yeast SMC Complexes by Chromatin Immunoprecipitation

A plethora of biological processes like gene transcription, DNA replication, DNA recombination, and chromosome segregation are mediated through protein–DNA interactions. A powerful method for investigating proteins within a native chromatin environment in the cell is chromatin immunoprecipitation (ChIP). Combined with the recent technological advancement in next generation sequencing, the ChIP assay can map the exact binding sites of a protein of interest across the entire genome. Here we describe astep-by step protocol for ChIP followed by library preparation for ChIP-seq from yeast cells.


Introduction
Chromatin immunoprecipitation (ChIP) is a powerful method for assaying protein-DNA binding in vivo and is broadly used to estimate the density of DNA-bound proteins at specific sites in the genome. ChIP is a multistep assay and every step needs to be optimized for consistent results. Briefly, protein-DNA associations are immobilized by cross-linking with formaldehyde [1][2][3] before shearing the chromatin, either mechanically [4] or by enzymatic digestion [5] into DNA fragments of average size 200-500 bp. Specific cross-linked protein-DNA complexes are then isolated by immunoprecipitation using an antibody to the protein of interest. Finally, the cross-links are reversed, and the retrieved DNA is analyzed to determine the sequences bound by the protein. ChIP followed by quantitative real-time PCR (ChIP-qPCR), using specific primers, can be used to measure protein association and relative abundance at a particular genomic locus. Alternatively, ChIP can be combined with next generation sequencing (ChIP-seq) to provide a genome-wide view of protein occupancy. While ChIP-seq allows for relative protein abundance at distinct chromosomal addresses to be compared within a sample, differences between samples cannot be quantified without introducing a method to normalize. Typically, this involves "spike in" of a known amount of DNA or cross-linked cells from a different species, with sufficient sequence divergence from the organism of interest to allow sequencing reads to be confidently distinguished bioinformatically [6][7][8]. This technique, referred to as calibrated ChIP-seq, makes it possible to quantitate genome-wide changes in the distribution of an epitope tagged protein and allows for quantification of differences in occupancy between experimental samples [8]. Calibrated ChIP-seq requires that both calibration and experimental organisms carry the same epitope tag and can be immunoprecipitated by the same protocol. For this protocol we use S. pombe to calibrate S. cerevisiae, a combination that also allows us to invert the roles, that is, calibrate S. pombe with S. cerevisiae.
The ChIP method described here has been optimized for use with chromatin from two species of yeast, S. cerevisiae and S. pombe; however, it should be easy to adapt it for use with other chromatin sources. To demonstrate the robustness of our ChIP and library preparation protocols we performed ChIP against the Scc1 subunit of the cohesin multiprotein complex, tagged with the 6HA epitope [9][10][11] . We have also used a similar protocol for the condensin subunit Brn1 [12] and for the meiotic counterpart of cohesin, Rec8 [13]. Here, we outline in detail an optimized protocol for crosslinking and harvesting cells, fragmenting chromatin, immunoprecipitating the desired protein-DNA complexes, and preparing the library for sequencing on the Illumina MiniSeq platform. A schematic stepwise representation of the method is illustrated in Fig. 1

Methods
Chromatin immunoprecipitation (ChIP) is broadly used to study chromatin dynamics. Changes in occupancy of chromosomal proteins at specific loci within the genome can be measured by using ChIP-qPCR. However, this technique is costly and time consuming with high variability per experiment. Alternatively, ChIP-seq can be used to measure differences in a protein's occupancy genome wide. Finally, calibrated ChIP-seq is essential when measuring changes in occupancy between different experimental samples.
Here we describe an optimized ChIP protocol for yeast SMC proteins that can be completed within 3 days for samples analyzed by qPCR and 4 days for samples to be further processed by calibrated deep sequencing. The protocol encompasses five distinct steps: cross-linking and cell harvesting; cell lysis and sonication; immunoprecipitation, decross-linking and DNA extraction and finally determination of the size and DNA concentration of sonicated samples. These five steps are outlined below.

Growth Conditions for SMC Proteins
S. cerevisiae strains for mitotic studies are grown in YPDA at 25 C. The most consistent results, at least for cohesin, are obtained when cells are arrested in metaphase of mitosis prior to the ChIP procedure. This can be achieved either by depletion of the anaphase-promoting complex subunit, Cdc20, or treatment of the cells with the microtubule-depolymerizing drug nocodazole. For depletion of Cdc20, we use a construct where CDC20 is under control of the methionine-repressible promoter, pMET3 (pMET3-CDC20). Briefly for Cdc20 depletion, dilute an overnight culture to OD 600 ¼ 0.2 in minimal media lacking methionine and grow for 1-2 h at 25 C to OD 600 ¼ 0.3-0.4. Dilute culture back to OD 600 ¼ 0.2 in same media and arrest cells in G1 by adding α-factor at 5 μg/ml for 1.5 h and at 2.5 μg/ml for an additional 1.5 h. Check microscopically that at least 90% of cells are arrested before collecting on a filter (Whatman ME25, 0.45 μm), washing with 10 volumes of medium lacking sugar with the aid of a vacuum pump. Quickly resuspend cells in YPDA containing 8 mM methionine and readd methionine to 4 mM every 45 min. Harvest cells after 2-2.5 h in a metaphase arrest confirmed by microscopy. For nocodazole arrest, dilute an overnight culture to OD 600 ¼ 0.2 in YPDA and grow for 1-2 h at 25 C to OD 600 ¼ 0.3-0.4. Dilute culture back to OD 600 ¼ 0.2 in YPDA media containing a mixture of nocodazole (15 μg/ml) and benomyl (30 μg/ml). Readd nocodazole every hour at 7.5 μg/ml. Harvest cells after 2-2.5 h confirming metaphase arrest by microscopy.
For studies of protein occupancy during meiosis we have used diploid S. cerevisiae strains of SK1 background including (a) REC8-3HA ndt80Δ (AM4015), as previously described [13] and (b) REC8-6HIS-3FLAG (AM11000). For inducing meiosis, SK1 strains are recovered from À80 C on YPG agar plates overnight at 30 C, before transferring to YPDA4% agar plates for a further 12-30 h at 30 C. Cultures are inoculated in liquid YPDA at 30 C with shaking for~24 h, prior to inoculating into BYTA medium to OD 600 ¼ 0.3 overnight. The next morning, cells are spun down, washed with dH 2 O and resuspended in SPO medium to OD 600 ¼ 1.8 and shaken at 30 C. For prophase I arrest (ndt80Δ) for Rec8 cells, 50 ml is harvested 6 h after resuspension in sporulation medium and the arrest is confirmed by FACS.
S. pombe strains used for calibration are listed in Subheading 2.1 and are grown in YES at 30 C.

Cross-Linking and Cell Harvesting
1. For ChIP-qPCR to measure the localization of the cohesin subunit, Scc1, in cycling cells, harvest 50 ml yeast cells of density 0.3-0.6 OD 600 grown in YPDA media. Alternatively, cells can be arrested in mitosis either by treatment with nocodazole or by depletion of Cdc20, as described above. For the less abundant cohesin loader subunit Scc2 and condensin subunit Brn1 harvest 100 ml yeast cells of density 0.3-0.6 OD 600 grown in YPDA. For the meiotic counterpart of cohesin, Rec8, harvest 50 ml yeast cells of density 1.8 OD 600 grown in SPO media. To cross-link cells, add 5 ml (11% formaldehyde in diluent buffer) to give a final concentration of 1% formaldehyde in the culture. Gently rotate on an orbital shaker at 90 rpm (with 1.8 cm orbit) at room temperature for 30 min for Scc1, Rec8, Scc2, and Brn1 (see Note 2).

For
ChIP-seq for the aforementioned proteins grow 2Â the amount of cell culture of yeast cells of density 0.3-0.6 OD 600 in YPDA media or 1.8 OD 600 in SPO media and process each 50 ml sample individually. Cross-link as in step 1. For calibrated ChIP-seq see Note 3.
3. To quench cross-linking, add glycine at a final concentration of 125 mM and incubate with gentle shaking for 5 min at room temperature (see Note 4).

Collect cells by centrifugation at 1800
Â g at 4 C.
5. Wash cells twice in 10 ml ice-cold TBS buffer and once in 10 ml ice-cold 1Â FA lysis buffer supplemented with 0.1% SDS.
6. Collect cells by centrifugation at 1800 Â g at 4 C.
7. Carefully aspirate the supernatant and snap freeze pellets in liquid nitrogen in fastprep screw-cap tubes. Store the pellets at À80 C until ready to use (PAUSE POINT).

Cell Lysis and Sonication
1. Thaw cells on ice. Add 1 volume (0.3-0.5 ml) of ice-cold 1Â FA lysis buffer supplemented with 0.5% SDS, 1 mM PMSF and protease inhibitors (Roche Complete EDTA-free tablet). For calibrated ChIP-seq see Note 3.  10. Store 10 μl of supernatant at À20 C. This will be the "Input" sample.
11. Use 100 μl of the chromatin preparation (step 9) to determine fragment size (see Subheading 3.5). 3. Place the tubes on the magnet and discard the supernatant. Perform the following washes, using 1 ml per sample of Wash buffer on a rotating wheel for 5 min at room temperature. Discard supernatants after each wash. 4. Following the final wash, place the samples on magnetic rack and discard the supernatant without disturbing the beads.

(a)
To reverse cross-linking and isolate DNA for qPCR, use Chelex 100 as previously described [14]. Add 0.2 ml 10% slurry (wt/vol) in sterile water Chelex-100 resin directly to the washed Dynabeads (IP sample) and to 10 μl of thawed "Input" samples (see  The sonicated chromatin samples (see Subheading 3.3, step 11) can be used to determine the fragment size.
1. To a 100 μl of Input sample add 80 μl of TE buffer containing 300 mM NaCl and decross-link at 65 C overnight.

Purify DNA using a PCR purification kit (see Subheading 2.2).
Run purified DNA on a 2% agarose gel with a 100 bp DNA ladder marker to determine fragment size. Ideally sonication should yield an enrichment of fragments between 200 and 400 bp (Fig. 2a).

ChIP-seq Library preparation
There are commercially available kits for generating DNA libraries but it is relatively straightforward and cost effective to create libraries using standard molecular biology reagents and custom oligonucleotides. This protocol can be completed within 1 day, and it comprises five distinct steps: blunting reaction; dA-Tailing to the 3 0 end of the DNA fragments; adapter ligation to the DNA fragments; PCR for enrichment of adapter modified DNA fragments; and library size selection. Finally, given the high cost of ChIP-seq runs and the time-intensive bioinformatics analysis and data validation, it is essential that the quality and the concentration of the libraries is validated by an Agilent Bioanalyzer (see Subheading 3.6.5, step 7) prior to sequencing. These five steps are outlined below.     7. Run the library on a Bioanalyzer to determine average fragment size and general purity. Use a High Sensitivity DNA Kit (Agilent Technologies) as per manufacturer's instructions.
Fragments of the sequencing library should have a size range of 150-300 bp (Fig. 2b, upper panel). If not pure, that is, adapter dimers are visible (Fig. 2b, lower panel), perform 1:1 AMPure purification to remove small adapter dimers.
8. Libraries are now ready for sequencing using a sequencing platform of choice.
9. The final concentration of the library to load on the flow cell is 1.5 pM with an Input-IP ratio 15%:85%. Perform paired-end sequencing with 76 bp-76 bp for Read1 and Read 2 (see Note 14).

AMPure Purification Protocol
AMPure purification relies on the principle of solid-phase reversible immobilization (SPRI) as previously described [16]. SPRI beads are paramagnetic and coated with carboxyl molecules, which reversibly bind DNA in the presence of polyethylene glycol (PEG) and salt (20% PEG, 2.5 M NaCl mix). PEG causes the negatively charged DNA to bind on the bead surface. This DNA immobilization is dependent on the concentration of PEG and salt in the reaction, and the volumetric ratio of SPRI beads to DNA is critical.
Equal volume of beads to DNA will give an SPRI -DNA ratio of one. As this ratio is changed the length of fragments binding and/or left in solution also changes. A lower SPRI-DNA ratio results in larger fragments at elution.
1. Equilibrate an aliquot for all purifications needed of AMPure XP beads at room temperature for 30 min before use. Vortex to resuspend.
2. Pipet carefully the indicated amounts so that no extra beads adhere to the outside of the tip.
3. Add the AMPure XP beads to DNA in solution and immediately mix thoroughly by repeated pipetting.
4. Incubate at room temperature for 10 min to allow binding of DNA to beads.
5. Place on a magnetic rack for 5 min.
6. Remove and discard the supernatant taking great care not to take any beads.
7. Keep sample on magnetic rack and add 250 μl of freshly prepared 80% ethanol without disturbing the beads.
8. Incubate for 30 s. Remove and discard all supernatant.
10. Let the beads air-dry for 2-3 min at room temperature (see Note 15).
11. Add the recommended amount of elution buffer (EB from Qiagen kit or ultrapure ddH 2 O) and resuspend the beads by pipetting.
12. Incubate at room temperature for 3 min.
13. Place in magnetic rack for 2 min.
14. Transfer the supernatant to a new DNA LoBind Eppendorf tube (if, for example, eluting in 30 μl, remove 28 μl very slowly, being careful to prevent bead carryover. If beads are accidentally removed, pipet the sample back into the tube and allow the beads to bind).

Bioinformatics Analysis
We carry out all data processing on the Ubuntu 16.04 (xenial) operating system. Basecalls are performed using Illumina Real-Time Analysis (RTA2) software on the MiniSeq System. FastQC is used to assess the quality of the raw sequence data (fastq reads), with fastq-screen used to detect any unwanted contamination. All quality control reports were aggregated with MultiQC [17]. ChIPseq paired end reads are trimmed with cutadapt, any adapter sequence is removed from the 3 0 end of reads using standard Illumina adapter sequences. Quality trimming is also performed from the 3 0 end using a user-defined cutoff (phred-33 quality 10).
After adapter and quality trimming, any read less than the defined minimum length (30 bp) is removed. Reads are mapped to both S. pombe calibration genome and S. cerevisiae w303 experimental genome, retaining only those reads that map to each reference.
To obtain reads mapping only to SacCer W303; trimmed fastq reads are first mapped with the MiniMap2 alignment tool [18] ("-ax sr" short genomic reads) to reference S. pombe, whereas unmapped S. pombe reads are selected using SAMtools [19] (include SAM Flag -F 4) and converted back into fastq format (interleaved), those unmapped S. pombe reads are then mapped to SacCer W303. Here, any unmapped reads are filtered out using samtools (exclude SAM Flag -F 4) and rDNA regions are removed from the section of chromosome XII which corresponds to the repetitive rDNA using BEDtools intersect [20], as this region is saturated with reads. To obtain reads mapping only to S. pombe the above process is performed in reverse. The original trimmed reads are also mapped to SacCer w303, unmapped SacCer w303 reads were selected using SAMtools, those unmapped SacCer w303 reads are then mapped to S. pombe, and unmapped reads are filtered out using SAMtools. Mitochondrial DNA is excluded using SAMtools for both genomes. In order to visualize mapped reads, bedGraphs are created from the aligned Binary Alignment Map (BAM) files using BEDtools genomeCoverageBed with reads per millions (RPM) normalization (calculated with custom script using SAMtools flagstat output) & UCSC wigToBigWig is used to convert these into BigWigs. For meiotic samples, where SK1 strains were used, mapping was performed to the SK1 genome, rather than SacCer3 as described above.
To generate the calibrated ChIP bigWigs; SAMtools flagstat is used to count reads mapping to SacCer3 w303 and S. pombe only for each sample, these values are then used to calculate the Occupancy Ratio (OR) value as previously described [8]; Wc*IPx/ Wx*IPc (W ¼ Input; IP ¼ chIP; c ¼ calibration genome (S. pombe); x ¼ experimental genome (sacCer w303)). Each OR value is used to calibrate ChIP bedgraphs using BEDtools geno-meCoverageBed and convert to bigWig with UCSC wigToBigWig. These bigWigs are viewable in a genome browser such as Integrative Genomics Viewer (IGV) [21] or the ensembl genome browser. All bigWigs from our published analyses are submitted to the Genome Expression Omnibus (GEO) archive and raw reads to the Sequence Read Archive (SRA).

Formaldehyde and PMSF are toxic if inhaled, ingested or
absorbed through the skin. Always wear a lab coat and gloves, and work in a chemical hood.
2. The cross-linking time and formaldehyde concentration can affect both the efficiency of chromatin shearing and of antigen precipitation. Shorter cross-linking times (5-10 min), lower formaldehyde concentrations (1%, wt/vol), or both, may improve shearing efficiency. However, for some proteins, especially those that do not directly bind DNA, this might reduce the efficiency of cross-linking and thus the yield of precipitated chromatin. It is advisable to perform a cross-linking time course to determine optimal fixation conditions. In vivo cross-linking for ChIP is traditionally achieved with formaldehyde; however, formaldehyde is a short spacer arm cross-linker (2 Å ), limiting its functionality. For higher order interactions, longer cross-linkers such as EGS (16.1 Å ) or DSG (7.7 Å ) or combination of cross-linkers can be used so to more efficiently trap larger protein complexes with complex quaternary structure [22].
3. For calibrated ChIP-seq use a 2:1 ratio of S. cerevisiae to S. pombe cells, as measured by OD 600 , mix pellets of different organisms in a single fastprep tube and lyse together as previously described [8]. Use the same batch of S. pombe in all samples of the same experiment. Perform each IP individually and pool samples together after the final wash step by combining beads from multiple IPs in the same 200 μl of TES elution buffer (see Subheading 3.4, step 4). Both calibration and experimental genomes need to express proteins tagged with the same epitope for immunoprecipitation and the calibration organism needs to be sufficiently similar that the ChIP protocol works for both.
4. Addition of glycine to quench the formaldehyde is particularly important when harvesting large volumes of cell culture as the harvesting process can be long and thereby can increase fixation time between samples.
5. Sonication conditions must be determined empirically for each cell type, and sonicator model; the optimal average DNA fragment size is below 0.5 kb. Overfragmentation of chromatin is not recommended as it can damage the protein epitope targeted by the antibody of choice. If the Diagenode water bath sonicator is not available, a probe sonicator or Covaris instrument can also be used. Sonication time and intensity will need to be optimized and DNA fragment size determined as in Subheading 3.5.
6. The amount of SDS in immunoprecipitation can interfere with antibody binding efficiency therefore, lower amount of SDS can also be used. Either omit SDS in 2xFA buffer or reduce SDS added afterward to a final concentration of 0.05% before chromatin immunoprecipitation.
7. The amount of antibody added should be in excess of the protein being immunoprecipitated and should be determined empirically. We use the following amounts of antibodies per IP: 10 μl 9E10 (Tonbo Biosciences) 7.5 μl 12CA5 (Roche), 10 μl SV5-PK1 (Bio-Rad), 10 μl anti-GFP (Roche), and 5 μl M2 FLAG (Sigma). No blocking and preclearing is required for the magnetic beads.
8. Variation in the amount of beads added can affect the specific signal-background ratio. Make sure to keep the slurry suspended while distributing.
9. To increase accuracy and reproducibility, we advise the use of an electronic dispensing pipette for qPCR. For primer design we recommend the use of Prime3Plus software. In the general settings of the software, we use product size range 70-200 bp (optimal amplicon size~120 bp), optimal primer size 20 bp in length and optimal primer Tm for use with Luna Universal Probe qPCR Master Mix should be designed to anneal at 60 C, with optimal primer GC% between 50% and 60%. Once primers are designed determine their efficiency and specificity using genomic DNA prior to performing qPCR with ChIP DNA sample. Finally, also include a No Template Control (NTC) in the reaction, where no amplification and no melting curve should be generated. 10. RNase treatment is important as high levels of RNA will interfere with DNA purification when using commercially available PCR purification kits. DNA yield can be markedly reduced as the columns become saturated.
11. For library preparation we recommend to use only filter-tips and 1.5 ml DNA LoBind tubes.
12. Always keep adapters on ice. The quantity of adapters stated here is recommended for 2 ng of DNA; however, the amount of adapters should be proportional to the amount of DNA used. Optimal concentration of adapters used is essential. High concentration can lead to adapter contamination in the final library, which can be visualized on the Bioanalyzer (Fig. 2b). We use for this protocol the NEXTflex DNA Barcodes-12 (Bioo Scientific; #NOVA-514102). The NEXTflex DNA Barcodes utilize an indexed adapter containing a 6 nt unique sequence. Details can be found in manufacturer's manual.
13. When PCR for library amplification is performed, minimal cycling is desirable. The fewer number of PCR cycles used to amplify libraries, the less biased the resulting libraries will be for the products that are more efficiently amplified. Overamplification can result in daisy-chains of fragments that can be visualized as a higher molecular weight peak on the Bioanalyzer. If the library amplification fails, more DNA template can be used. For IP samples use up to 20 μl of template DNA.
14. Several different next generation sequencers are available; for this protocol we use the Illumina platform. While platforms vary by target sequence length, accuracy and cost all give reproducibly comparable results. In order to perform experiments in a cost-effective manner, multiplexing can be used, that is, multiple ChIP-seq libraries, each carrying a different barcode to identify different samples can be sequenced together on a single flow cell of MiniSeq or lane of a Hiseq. The output of the Illumina MiniSeq is~25 M clusters. We typically, sequence 8-10 barcoded uncalibrated samples in a single pooled library or 4-5 barcoded calibrated samples in a single pooled library. Typically one ChIP library generates six to ten million reads [23,24]. However, the above is subject on the level of enrichment of the protein of interest and the resolution required.
15. Overdrying the AMPure beads after the washing step will negatively impact on the DNA recovery. Beads are dry enough as soon as they lose their sheen.