A method for assessing histone surface accessibility genome-wide

Abstract

The assembly of DNA into nucleosomes and higher order chromatin structures serves not only as a means of compaction but also organizes the genome to facilitate crucial processes such as cell division and regulation of gene expression. Chromatin structure generally limits access to DNA, with the accessibility of DNA in chromatin being regulated through post translational modification of the histone proteins as well as the activity of chromatin remodeling proteins and architectural chromatin factors. There is great interest in assessing chromatin accessibility genome-wide to identify functional elements associated with enhancers, promoters, and other discontinuities in the compacted chromatin structure associated with gene expression. As the vast majority of techniques rely upon assessment of the exposure of the underlying DNA, we describe here a general method that can be used to assess exposure of internal and external histone protein surfaces. We demonstrate the feasibility of this method, in the organism S. cerevisiae. Our method relies on substitution of residues residing on selected histone protein surfaces with cysteine, and assessment of exposure by reaction with a thiol specific reagent, biotin-maleimide. We demonstrate that modified nucleosomes can be efficiently excised from nuclei treated with the reagent via a one-step purification process. After library preparation and deep sequencing, selected nucleosomes are typically ~25-fold enriched over background signals and exhibit phasing with respect to transcription start sites in yeast that is identical to an unselected population.

1. Introduction

Chromatin is comprised of a basic repeating subunit known as the nucleosome. Each nucleosome contains two copies each of four unique histone proteins: H2A, H2B, H3, and H4, which form a protein octamer. This octamer is wrapped by ~1.65 left-handed turns of ~147 bp of DNA, forming a nucleosome core [1]. Nucleosome cores are separated by ~10 to 90 bp stretches of linker DNA that is not strongly associated with the core histones. In higher eukaryotes, an additional ‘linker’ histone binds in a structure-specific manner to the surface of the nucleosome core and also associates with the linker DNA [2]. Linker histones stabilize folding of strings of nucleosomes into higher-order chromatin structures (see below).

Native chromatin and reconstituted oligonucleosomes undergo folding and compaction transitions dependent on solution ionic strength and specific cations. In low ionic strength solutions, strings of nucleosomes exist in an extended “beads on a string” structure [3]. In solutions containing physiological ionic strengths and multivalent cations oligonucleosomes condense, forming higher order structures depenedent upon inter-nucleosome interactions. Perhaps the most studied of these higher order structures is the 30 nm chromatin fiber, whose exact formation and existence in vivo is an area of debate[4–6].

Chromatin structure can be modulated by histone posttranslational modifications and the binding of chromatin factors that either promote or restrict chromatin condensation. As a result, chromatin is a highly dynamic entity containing regions of lower degrees of folding and condensation as well as more condensed and compact structures. Open portions of the genome tend to be comprised of actively expressed genes, promoters and enhancers. Conversely, closed regions of the genome tend to contain repressed genes, or be gene poor, comprising intergenic regions, heterochromatic, centromeric and telomeric regions [7]. Because alterations in the chromatin structure have important consequences, understanding where different regions of chromatin are located and how this dynamic process is modulated has been an area of tremendous interest.

The development of next generation sequencing (NGS) coupled with chromatin immunoprecipitation (ChIP) for histones, post-translational modifications and other chromatin binding proteins have allowed for genome-wide annotation of multiple aspects of primary chromatin structure. Some of these aspects are nucleosome positioning, chromatin domain identification, and nonhistone protein binding sites within the genome [8–11]. Chromatin crosslinking and capture techniques have provided information on long-range chromatin interactions [12–15]. While informative, these techniques do have limitations, namely the reliance upon restriction enzymes whose sites are heterogeneously spread across a genome and and issues with crosslinking and ligation efficiencies, that reduce effectiveness of mapping at short and mid-range genomic distances [16].

Early techniques such as DNase and MNase hypersensitivity revealed that specific regions near active genes were sensitive to nuclease digestion [17–19]. Additionally, genes which were impervious to enzymatic digestion under non-induced conditions, became susceptible to digestion in response to heat stress induction, indicating the presence of different chromatin folding states associated with gene activation [17–19]. Evidence for structural differences between euchromatin and heterochromatin has also been inferred by the analysis of chromatin derived from chicken erythrocytes. Separation of the active and inactive ß-globin locus on sucrose gradients resulted in faster migration of the inactive locus, indicating a more condensed, folded chromatin structure [20, 21]. Genome-wide approaches, including mapping DNase I hypersensitive sites and ATAC-Seq provide genome-wide maps of sites accessible to DNA-dependent enzymes. ATAC-seq, relies on the integration of adaptors into the genomic DNA via Tn5 transposase and has been used to identify accessible (i.e. protein poor) regions of chromatin [22]. Additionally, the positioning of nucleosomes and other DNA binding protein have been mapped [22, 23]. This method also has the advantage of requiring significantly fewer cells than the more traditional DNaseI and MNase based techniques and has even been applied to single cell analysis [22, 24]. However, like other methods, the results of ATAC-seq can be influenced by the preference of Tn5 for G/C sequences [25, 26].

These efforts have greatly expanded our understanding of the regulation of chromatin accessibility, but they have primarily explored DNA accessibility. Although DNA accessibility ultimately impacts the expression of genes, assaying solely for this characteristic provides incomplete information with respect to higher order chromatin structures. We have developed a method to report on the genome-wide accessibility of histone surfaces in the nucleosome. We demonstrate proof-of-principle for this method by probing accessibility of an external nucleosome surface in S. cerevisiae, based on the accessibility of a single cysteine residue exposed on the flat protein surface of the nucleosome to a thiol specific reagent (Figure 1). Application of this technique to both external and internal protein surfaces in the nucleosome offers great potential to better understand chromatin structure and dynamics.

Fig. 1:

Schematic of method for accessibility of histone surfaces in native yeast chromatin. A. A rationally selected histone residue is substituted for cysteine (H2B S116C in this work, red residue). B, Nuclei containing native yeast chromatin with introduced cysteine are prepared and treated with biotin-maleimide (BM, red chevron). Mononucleosomes are then generated by digesting chromatin with micrococcal nuclease, and incubated with streptavidin-agarose resin to affinity purify BM-modified nucleosomes. The DNA recovered from BM-modified nucleosomes is sequenced and mapped to S. cerevisiae reference genome (SacCer2).

2. Methods

In order to interrogate chromatin structure in vivo through the accessibility of discrete histone surfaces, we developed a novel biotin-maleimide-native nucleosome affinity-purification based technique using S. cerevisiae as a model system. None of the four core histones nor the two histone variants, H2A.Z and Cse4 present in S. cerevisiae contain cysteine residues. Therefore one may introduce unique cysteine substitutions at rationally selected sites within the core histones, exposed on specific external or internal surfaces within the nucleosome. The cysteine thiol will preferentially react with maleimide-based reagents, according to the accessbility of the thiol group [27]. We reasoned that biotin-maleimide (BM) could be used to selectively identify nucleosomes with the specific protein surfaces exposed. This hypothesis is supported by previous in vitro work which found that accessibility of various sites within the H2B N-terminal tail could be probed with a similar reagent, fluorescein 5-maleimide [27].

To test this idea we created yeast cells in which all H2B was replaced with H2B S116C. This residue is exposed on the protein surface of nucleosomes, near the acidic patch (Fig. 1A), and predicted to be occluded in models of nucleosome-nucleosome stacking in higher order chromatin structure (Fig.1B) [28]. Indeed, in vitro studies with model nucleosome arrays indicate that accessibility of cysteine at this position in H2B is reduced several-fold in condensed structures (unpublished results). Such packing and reduced accessibility may occur in regions of heterochromatin such as is found at the mating-type loci and telomeres in yeast cells. Nuclei from wild-type and H2B S116C strains were isolated and treated with biotin-maleimide (BM). The maleimide reaction was quenched with the reducing agent ß-mercaptoethanol. Following BM modification, the chromatin was digested with micrococcal nuclease (MNase) such that mononucleosomes were the predominant species. The digested chromatin was then incubated with streptavidin linked agarose to affinity purify nucleosomes that had reacted with BM. The DNA from the affinity purified nucleosomes was recovered, sequenced and aligned to the reference genome of S. cerevisiae to identify and map nucleosomes with the accessible histone surface.

2.1. Preparation of S. cerevisiae strains

Strain for H2B mutant H2B S116C: YDC417 (derived from W303 strain)

Genotype: W303 MATa can1–100 his3–11 leu2–3,112 lys2delta trp1–1 ura3–1 GAL+ ADE2+ hta2-htb2::TRP1

Overview: The strain YDC417 can be used for generating mutants in the H2A (HTA1) and H2B (HTB1) histones, and is derived from the strain W303 [29]. One of the two loci for H2A and H2B, the HTA2-HTB2 locus, was replaced with a functional TRP1 allele eliminating H2A and H2B production from this locus. In order to introduce the desired mutation in the HTA1-HTB1 locus, an integration plasmid, p592, is used as described in [30]. Briefly this plasmid contains the HTA1-HTB1 locus as well as a functional copy of the HIS3 allele which will confer growth on histidine deficient media. The serine to cysteine mutation is achieved in this plasmid using Q5 site-directed mutagenesis (NEB E50554S). The plasmid was sequenced to confirm that the H2B S116C mutation was present. In order to transform the YDC417 yeast strain, the mutated and wild-type plasmids are digested with specific restriction enzymes, BamHI-HF (NEB R3136S) and EcoRI-HF (NEB R3101S), to release a linear DNA fragment containing these three genes (Fig. 2). Transformation of this fragment into the YDC417 strain results in integration into the genome via homologous recombination (Fig. 2). Transformants are plated on media deficient in histidine to select for cells that have integrated the linearized HTA1-HTB1-HIS3 insert from p592. To determine if homologous recombination of this fragment occurred in the proper orientation and that the desired mutation is present, genomic DNA is amplified by PCR and fragments from the region of interest are sequenced. It is important to note that because only one locus contains an H2B gene (HTB1) all H2B produced in the mutant strain will be generated from the transgene.

Fig. 2:

Insertion and confirmation of proper integration and cysteine mutation. Genomic DNA from colonies that exhibited growth on SC-HIS media was extracted and analyzed for proper integration and desired mutation. Primers were designed such that one was specific to a region of the genomic DNA (red line) just outside the locus of interest (purple arrow) and the other was specific to the plasmid (blue line), specifically a region of the HIS3 gene (green arrow). If integration was successful, a 3.2 kb fragment should be generated. Strains where this 3.2 kb fragment was identified were then sequenced to confirm the H2B S116C mutation.

2.1.1. Linear Transformation Protocol

1. Inoculate a 5 mL YPD culture with YDC417 cells and grow overnight at 30°C with shaking (~225 rpm).

2. Dilute overnight culture into 50 mL YPD such that OD₆₀₀ is ~0.1 and grow at 30°C with shaking (~225 rpm) until the OD₆₀₀ is ~0.5.

3. 10 mL of culture is placed in 50 mL Falcon tube and centrifuged at 1 kG (Sorvall RC6+, 3000 rpm for 5 min with F21–8×50y rotor at 4°C).

4. Supernatant is removed and the cell pellet is washed 2x with 5 mL of 100 mM LiAc.

   1. Repeat centrifugation conditions from previous step.

5. Resuspend cell pellet in 100 μL of 100 mM LiAc.

6. Label 1.7 mL Eppendorf tubes and aliquot 10 μL of 10 μg/μL salmon sperm DNA along with 500 ng of appropriate template.

7. Add cells from step 5 to each tube and incubate at 30°C for 15 min.

8. After 15 min incubation at 30°C, add 600 μL of LiAc PEG (100 mM LiAc, 48% PEG) solution to each tube and mix well by inversion.

9. Incubate tubes at 30°C for 30 min.

10. Add 68 μL DMSO to each tube and mix by inversion.

11. Heat shock cells at 42°C for 15 min.

12. Harvest cells via centrifugation (Eppendorf 5415C centrifuge; 1500 rpm for 7 min).

13. Remove supernatant and add 1 mL of YPD to each tube.

14. Shake tubes for 4 hours at 30°C with loose cap for aeration.

15. Plate 250 μL of cells on SC-HIS plates and incubate at 30°C for 2–3 days (until colonies are easily visible).

16. Store plates at 4°C.

2.1.2. Isolating genomic DNA from yeast and sequencing to confirm proper integration

1. Transfer 2 mL of culture to Eppendorf tubes, 1 mL in each tube and centrifuge at 15,000 RPM at 4°C (Eppendorf 5415C).

2. Remove supernatant and add 200 μL of genomic DNA prep buffer (2% Triton X-100, 1% SDS, 100 mM NaCl, 10 mM Tris pH 8.0, 1 mM EDTA pH 8.0), 300 μL glass beads (Sigma Aldrich Z250473) and 200 μL phenol-chloroform isoamyl alcohol pH 8.0 and vortex at top speed for 3 min.

3. Centrifuge sample at 15,000 RPM for 1 min at RT and remove top layer and place in new tube.

4. Add 200 μL TE and spin for 5 min 15,000 RPM at RT.

5. Transfer aqueous layer ( ~400 μL) to new tube.

6. Add 1 mL of 100 % EtOH and place sample on dry ice for 15 min and then centrifuge at 15,000 RPM for 5 min at 4°C.

7. Remove supernatant and resuspend in 400 μL TE and add 1 μL RNaseA (10 μg/μL).

8. Dissolve for 5 minutes by vortexing at medium speed and then add 1 mL 100 % EtOH, 10 μL 3 M NaOAc pH 5.2 and place on dry ice for 15 min.

9. Centrifuge for 15 min at 15,000 RPM at 4°C and remove supernatant.

10. Resuspend in 50 μL of TE.

11. DNA is sequenced using primers that are specific to plasmid and genomic DNA to ensure properintegration (Fig. 2).

• Genomic DNA primer:

• 5’ GCGGACTCGAAGTAGAGAAGGCATCC 3’

• Plasmid insert DNA primer:

• 5’ CGTCTATGTGTAAGTCACCAATGCACTCAACG 3’

2.2. Preparation of yeast nuclei and biotin-maleimide modification of yeast chromatin

Overview: Wild-type and mutant (H2B S116C) yeast strains are grown to mid-log phase (OD₆₀₀ is 0.5–0.8) and harvested. Spheroplasts are generated from these cells via lyticase digestion. Spheroplasts are lysed and nuclei are isolated at which point biotin-maleimide is introduced to probe for accessible H2B S116C in native chromatin. After quenching the maleimide reaction with ß-mercaptoethanol, the modified nuclei are subjected to MNase digestion such that the vast majority of chromatin is digested to core particle size (~149 bp). Digestions which achieved this were then used for streptavidin-based affinity purification to isolate nucleosomes which had become modified by biotin-maleimide.

1. Colonies of appropriate yeast strain(s) are selected from plates and grown overnight in 5 mL SC-mediaat 30°C with shaking (~225 rpm).

2. A 1:10 dilution of overnight culture(s) is prepared and the OD₆₀₀ determined.

3. A 250 mL culture is inoculated with the overnight culture such that the initial OD₆₀₀ is ~0.10. 250 mL cultures are grown at 30°C with shaking (~225 rpm) until OD₆₀₀ is 0.5–0.8 (mid-log phase).

4. Cells are harvested via centrifugation (Sorvall RC6+ centrifuge, 3000 rpm; F12–6×500 rotor; 4°C). Cells are resuspended in 250 mL ddH₂O and the spin is repeated. Cell pellets are stored at −80°C.

5. On the day of nuclei preparation, cells are thawed on ice and nuclei are prepared according to [31] with minor alterations described below.

6. Spheroplasts are generated as per [31] by adding 6000 U of high-grade lyticase (Sigma-Aldrich L2524).

7. Nuclear pellet is then resuspended in 2.3 mL of MNase digestion buffer without CaCl₂ and ß-mercaptoethanol.

8. EZ-Link™ Maleimide-PEG2-Biotin (ThermoFisher Scientific 21901BID) is dissolved in DMSO to make a 200 mM biotin-maleimide (BM) stock solution. BM stock is diluted in Mnase digestion buffer lacking both CaCl₂ and ß-mercaptoethanol to a final concentration of 0.6 mM and a final volume of 100 μL.

9. The 100 μl of 0.6 mM BM is added to 2.3 mL of nuclei, mixed and then incubated on ice for 1 min.

10. The maleimide reaction is stopped by the addition of ß-mercaptoethanol (Sigma Aldrich M6250) to a final concentration of ~5 mM. One μL of ~14 M ß-mercaptoethanol is added to 99 μL of MNase digestion buffer, mixed and added to 2.3 mL of nuclei and mixed by light flicking.

11. After one minute incubation on ice, 20 μL of MNase digestion buffer supplemented with 62.5 μM CaCl₂ is added to a final concentration of 0.5 mM, and the solution mixed by light flicking.

12. ~400 μL of nuclei are then aliquoted into six 1.7 mL Eppendorf tubes and incubated at RT for ~ 2 min. Each of the 6 tubes are treated with an increasing amount of MNase such that at least one of the enzyme concentrations will yield predominantly mononucleosomes (~150 bp).

13. Micrococcal nuclease (Worthington Biochemical; LS004798) is prepared as in [28] to (10 units/μl). MNase is then added in the appropriate amounts and samples are mixed by flicking. Chromatin is digested for 3 minutes at RT. Digestion is stopped by the addition of 100 μL of cold MNase stop solution (5 mM EDTA, 5 mM EGTA, 0.05% NP-40).

14. Samples are incubated on ice for ~ 1 min and then centrifuged at 13,000 RPM for 1 min in an Eppendorf 5415C centrifuge at 4°C.

15. Supernatant is recovered (~500 μL) and applied in its entirety to a 0.45 μM microfuge tube filter units (FisherScientific UFC30HV00) and the centrifugation repeated as in previous step.

16. 100 μL of recovered flow-through is removed for input DNA analysis (2.2.1).

17. The remaining ~400 μL is frozen on dry ice then stored at −80°C for affinity purification.

18. Input DNA is prepared as in [31] with volumes adjusted for 100 μL of sample.

2.2.1. Analysis of MNase digestions and input DNA preparation

The DNA from the 100 μL of recovered flow-through from 2.2 step 17 is purified in order to identify the MNase digestion conditions that yielded the desired distribution of chromatin subunits. For the purposes of these experiments we sought a digestion profile that yielded predominantly mononucleosomes.

1. 100 μL of digested chromatin (step 17, section 2.2) is treated with 5 μL of 20% SDS and 25 μL of 5 M K-acetate and vortexed.

2. 125 μL of chloroform is added and the sample is vortexed again and then centrifuged at 15,000 RPM at RT in Eppendorf 5415C.

3. The supernatant is removed and placed in a new tube with 125 μL of chloroform, vortexed and then centrifuged as in step 2.

4. The supernatant (~125 μL) is recovered and placed in new Eppendorf tube with 90 μL of isopropanol (~70% volume of supernatant) and 1 μL of 10 μg/μL glycogen. The sample is vortexed and then stored at −80°C for 1 h.

5. The sample is then centrifuged for 30 min at 15,000 RPM in Eppendorf 5415C at RT. The supernatant is removed and the pellet is washed with 500 μL cold 70% EtOH, and centrifuged for 5 min at 15,000 RPM for 5 min at RT in Eppendorf 5415C.

6. Supernatant is removed and pellet is allowed to airdry for ~5 min at RT. The pellet is then resuspended in 50 μL of TE. The resuspended pellet is then treated with 1 μL of 10 μg/μL RNaseA and incubated at RT overnight.

7. After RNaseA treatment, 10 μL of sample is analyzed via agarose gel (2% agarose; 0.5X TBE; 0.04% SDS) in order to assess digestion profile of samples. Digestion(s) that have yielded predominantly mononucleosomes (Fig. 3A) are diluted 10-fold and stored at −80°C until DNA library preparation.

8. The corresponding sample of digested chromatin (step 18, section 2.2) is then used for affinity purification.

Fig. 3:

A. Titration of micrococcal nuclease to biotin-maleimide modified yeast chromatin to identify digestion conditions that yield predominantly mononucleosomes. Biotin-maleimide modified nuclei (~2.4 mL) are subqaliquoted into 6 400 μL reactions and treated with 30, 60, 120, 240, 480 or 960 units of micrococcal nuclease. DNA is purified using ~1/5^th of digestion. DNA is then analyzed via agarose gel electrophoresis (2% agarose, 0.5X TBE, 0.04% SDS) to determine the extent of digestion for each enzyme condition. Digestions which yield predominantly mononucleosomes (denoted by gold star) are selected for affinity purification. B. Western blots of soluble chromatin from wild-type and H2B S116C nuclei. Blots were performed with anti-yeast H2A (left) and streptavidin-HRP, as described in the text. Note an unknown protein (X) runs just above the putative yH2B band,

2.2.2. Streptavidin agarose affinity purification of biotin-maleimide modified nucleosomes

1. Optimally MNase digested chromatin samples are removed from the −80°C and thawed on ice.

2. While chromatin is thawing, 150 μL of streptavidin-agarose slurry (Solulink N-1000–002) is placed into 1.7 mL Eppendorf tubes for each chromatin sample to be subjected to affinity purification.

3. The slurry is washed once with 1 mL of 1X PBS pH 7.4. The resin is centrifuged for 3 min at 3,000 RPM (Eppendorf 5414C) and the supernatant is discarded.

4. The streptavidin-agarose resin is then washed three times with 0.5 mL of 1X PBS pH 7.4 supplemented with 5 mg/mL BSA (NEB B9000S), with centrifugation performed as for the previous step after each wash.

5. 200 μL of each chromatin sample is then added to tubes containing the washed streptavidin-agarose resin in 800 μL of AP binding buffer (10 mM Tris pH 8.0, 1 mM EGTA, 0.05% NP-40, 150 mM NaCl, protease inhibitors (Sigma Aldrich 11873580001). Note that buffer is stored at 4°C. Sample is then incubated with streptavidin-agarose for 30 min at 4°C with gentle tumbling.

6. After incubation, the sample is centrifuged as in step 3. The supernatant is recovered and stored at −80°C for Western blotting analysis. Streptavidin agarose is then washed once with 1 mL of AP-binding buffer for 10 min at 4°C with gentle tumbling.

7. The streptavidin-agarose is centrifuged as in step 3, the supernatant is discarded, and two high salt washes are performed (AP-binding buffer with 600 mM NaCl). The washes are each 1 mL and performed in the same manner as first wash.

8. Washed streptavidin-agarose is then treated with 150 μL of elution buffer 1 (1% SDS, 140 mM NaCl, 50 mM Tris pH 8.0, 10 mM EDTA pH 8.0) for 10 min at RT with occasional manual agitation.

9. The resin is centrifuged as in step 3 and the ~150 μL of elution buffer 1 is recovered and placed in a new 1.7 mL Eppendorf tube on ice. The streptavidin-agarose resin is then treated with 100 μL of elution buffer 2 (elution buffer 1 with 0.67% SDS) for 10 min at RT with occasional agitation, then centrifuged again and ~100 μL of elution buffer 2 recovered and added to the first elution to generate ~250 μL total eluate.

10. To the eluate, 62.5 μL of 5 M potassium-acetate is added along with 310 μL of chloroform. The sample is vortexed and then centrifuged for 5 min at 15,000 RPM in a microfuge (Eppendorf 5415C).

11. The supernatant is recovered, placed in a new Eppendorf tube and mixed with 310 μL of chloroform, vortexed and the sample is centrifuged as in the prior step.

12. The supernatant (~300 μL) is recovered and placed in a new Eppendorf tube with 210 μL of isopropanol (~70% volume of supernatant) and 1 μL of 10 μg/ μL glycogen. The sample is vortexed and stored at −80°C for 1 h.

13. The sample is then centrifuged at 15,000 RPM for 30 min at RT in the microfuge. The supernatant is removed and pellet is washed with 500 μL ice cold 70% EtOH, and centrifuged again for 5 min at 15,000 RPM for 5 min at RT in the microfuge.

14. The supernatant is removed and pellet is allowed to airdry for ~5 min at RT. Pellet is then resuspended in 50 μL of TE. Resuspended pellet is then treated with 1 μL of 10 μg/ μL RNaseA and incubated at RT.

2.3. Testing selective retention of cysteine bearing nucleosomes in soluble yeast chromatin preparations

While the introduced cysteine is the only cysteine present in any of the S. cerevisiae histones, there are numerous nuclear proteins with accessible cysteine residues. Indeed, a Western blot of an SDS-PAGE of nuclear proteins probed with streptavidin-HRP shows that such proteins greatly outnumber core histones (Fig. 3B). Given that a portion of these proteins likely interact with chromatin, we devised an in vitro system for simple tracking of nucleosomes during the affinity purification process, which allowed us to determine whether H2B S116C-BM nucleosomes (or any other modified nucleosomes) are selectively retained through binding to the streptavidin-agarose resin.

To generate radiolabeled DNA that would be impervious to phosphatases in the yeast nuclear preps, we prepared DNA templates containing internal radioactive label for nucleosome reconstitutions. As described below, nucleosomes were reconstituted with Xenopus laevis WT histones, or with nucleosomes containing H2A/H2B S112C quantitatively modified with biotin-maleimide. Note that H2B 112 in X. laevis was chosen because it is the orthologous position to H2B 116 in S. cerevisiae H2B. Wild-type and H2B S116C- BM radiolabeled nucleosomes were introduced into MNase digested chromatin supernatants prepared from unmodified and BM modified yeast chromatin so that we could assess nonspecific retention and determine the extent of recovery for BM modified nucleosomes in these complicated preparations.

2.3.1. Preparation of nucleosomes containing internally radiolabeled DNA and either unmodified H2B or H2B S112C modified with biotin-maleimide

1. PCR primers (Forward: 5’- CACAGGATGTATATATCTGAC −3’; Reverse: 5’ CCTGGAGAATCCCGGTGCCG −3’) were chosen to generate a 149 bp DNA fragment from the 601 nucleosome positioning template found in the plasmid CP1024 [32]. Primers were diluted to 1 μg/μL concentration and 0.5 μL of each was added.

2. PCR was carried out with [α−32P]dCTP (PerkinElmer BLU013Z250UC) and dATP, dGTP, dTTP and dCTP. dATP, dGTP and dCTP were added such that their final concentration was 2.5 mM. Additionally, dCTP was added to a final concentration of 2.5 μM along with 10 μL of [α−32P]dCTP (6000 Ci/mmol). PCR volume was 50 μL:

   • 1
     94°C for 1 min 2.
   • 2
     1. 94°C for 45 s
     2. 52°C for 1 min
     3. 72°C for 1 min
   Repeat step two 25x

   • 3
     72°C for 10 min
   • 4
     Hold at 4°C

   Nucleosomes were reconstituted with the internally radiolabeled 149 bp 601 DNA fragment, 4.5 μg cold pBS plasmid, 2.5 μg of either X.laevis H2A/H2B or H2A/H2B S112C dimers and 2.5 μg H3/H4 tetramers (prepared from G.gallus erythrocytes) by standard salt dialysis as described previously [33]. Final dialysis was into 2 L of TE overnight.

3. Radiolabeled nucleosome reconstitutions were then subjected to a 7% – 20% sucrose gradient (w/v in TE). Samples were then centrifuged for ~18 h at 34,000 rpm in Beckman Coulter Optima L-90K Ultracentrifuge using SW-41 rotor.

2.3.2. Analysis of Specific Retention of BM-modified H2B S112C nucleosomes

Specific radiolabeled nucleosomes containing unmodified and BM modified Xenopus H2B were introduced into the clarified soluble chromatin preparations from WT yeast cells, immediately after stopping MNase digestion (after Step 15, Section 2.2). The following combinations were investigated:

1. Radiolabeled nucleosomes containing unmodified WT histones were introduced into BM modified yeast chromatin.

2. Radiolabeled nucleosomes containing BM modified H2B S116C were introduced into unmodified yeast chromatin.

3. Radiolabeled nucleosomes containing BM modified H2B S116C were introduced into BM modified yeast chromatin.

Chromatin mixtures were then incubated with streptavidin-agarose and radiolabeled nucleosomes were tracked via autoradiography. Specifically, 200 μL reconstituted, sucrose gradient-purified nucleosomes (~200,000 cpm) was mixed with 200 μL of yeast chromatin and the volume adjusted to 1 mL with AP binding buffer. Affinity purification was carried out as described in section 2.2.2 with the exception of exchanging one of the 600 mM NaCl washes for a 35 mM NaCl wash. The column flow-through, all washes and elutions were recovered for analysis. Samples of each were loaded onto a native gel (0.7% agarose; 0.5X TBE) and run at 100 V for 45–60 min. Gel was dried and exposed to phophorimage screen overnight and imaged on a Typhoon (GE Healthcare Life Sciences).

2.4. Preparation of DNA libraries and analysis of sequenced reads

2.4.1. DNA library generation

DNA libraries for input and affinity purified samples are prepared using NEBNext® Ultra™ DNA Library Prep Kit for Illumina® (New England BioLabs E7370) and NEBNext® Mutiplex Oligos for Illumina® (New England BioLabs E7335). No alterations were made to New England BioLabs protocol for kits. Briefly, the digested DNA fragments are repaired and then the adaptors are ligated to the repaired DNAs. Prior to PCR, the adaptor ligated DNAs are subjected to a cleanup step where we do not employ a size selection. For the experiments described in this manuscript, we find that using 5 μL (~1/3) of the adaptor ligated DNA in the PCR amplification along with 10 iterations of the cycling conditions yields sufficient material for NGS sequencing. In order to determine fold-enrichment of cysteine mutant nucleosomes with respect to wild-type, we added 1 ng of mouse nucleosomal DNA to all samples just before proceeding with library preparation. Paired end reads are aligned to respective reference genomes for yeast and mouse DNAs (SacCer2 and mm10 respectively) using Bowtie2. The number of aligned yeast and mouse reads are determined for each sample and the ratio of reads is calculated. For example, if there were 10 million total yeast reads and 5 million mouse reads for a given sample, the ratio would be 2. This ratio is determined for wild-type and mutant input and affinity purified (AP) samples. The AP/Input ratio is then determined for each by dividing the yeast/mouse DNA ratio of the affinity purification by that of the input. To determine the overall fold-enrichment (i.e. mutant vs. WT), the AP/input ratio of the mutant is divided by that of the wild-type.

3. Results and Discussion

We describe a method for interrogating chromatin structure on the basis of a rationally selected amino acid residue in a single core histone in native chromatin. The accessibility of the residue is gauged by the extent of modification with biotin-maleimide and modified nucleosomes affinity purified and sequenced. Reads from modified nucleosomes are then mapped to the reference genome using well characterized MNase-seq and ChIP-seq methodologies. Once the mutant strain was generated we conducted growth and mating assays with the wild-type and mutant strains. We observed there were no differences in growth under optimal conditions (YPD; 30°C). Furthermore, growth was similar between strains at both 20°C and 37°C, indicating that the cells were not markedly affected by the S116C mutation in H2B. In order to investigate this further, we conducted mating assays with wild-type and mutant strains as yeast that are unable to properly silence the a or α mating loci via heterochromatin formation have a compromised capacity to mate with a strain of the opposite mating type. We found that the wild-type and H2B S116C mutant strain had comparable levels of diploids suggesting that heterochromatin function was not impacted by the mutation. Lastly, a comprehensive analysis performed in S. cerevisiae did not reveal any defects upon mutating H2B S116 to alanaine [34]. Finally, the WT and S116C input nucleosomes show identical nucleosome phasing maps, indicating that the overall chromatin organization in the cells harboring the H2B S116C mutation is normal (see below).

Another major concern was that there would likely be a large number of cysteine-containing nonhistone proteins with exposed thiols in the nuclei that would be modified with BM. To determine whether these nonhistone proteins could promote retention of unmodified nucleosomes, or hinder selective retention of modified nucleosomes. To address this, we spiked in reconstituted nucleosomes containing radiolabeled 147 bp DNA and unmodified H2B or H2B S112C quantitatively modified with BM and tracked them through the affinity purification process.

We first confirmed that nucleosome cores containing H2B S112C modified with BM were specifically retained on the SA-agarose resin. We found radiolabeled unmodified nucleosomes mixed with soluble yeast chromatin derived from nuclei treated with BM elute primarily in the flow through and wash fractions, and not are not detectable in the elution fraction (Fig. 4A). These results indicate that BM-modified non-histone nuclear proteins do not bind and retain the unmodified nucleosome cores through the high-salt wash steps in the procedure. In contrast, radiolabeled nucleosomes containing BM-modified H2B S112C were retained and appear in the elution fraction from the resin, indicating the modified site (H2B residue 112 in X. laevis and 116 in S. cerevisiae) is available for binding by the streptavidin and that nucleosomes containing this modified residue are specifically retained over unmodified nucleosomes.

Fig, 4.

Selective retention of nucleosomes containing biotin-maleimide (BM) modified H2B S112C from soluble yeast chromatin. Nucleosomes containing an internally 149 bp labeled DNA fragment and X. laevis wild-type H2B or H2B S112C modified with BM were introduced into soluble MNase-digested yeast chromatin prepared from nuclei that were either unmodified or modified with BM. Samples were processed as described in the text, run on native 0.7% agarose gels, the gels dried, and phophorimaged. A. Specific retention of nucleosomes containing biotin-modified H2B S112C. Modified (lanes 1–7 and 9–21) or unmodified (lanes 8–14) nucleosomes were incubated in chromatin extracts from unmodified or BM modified nuclei (lanes 1–7 or lanes 8–21, respectively). Note elution fractions contained SDS, resulting in formation of free DNA. B. Saturation of SA-agarose blocks nucleosome retention. As in A except with increasing amounts of SA-agarose resin.

We also wondered whether the binding capacity of the resin could affect specific retention of modified nucleosomes. Given the advertised binding capacity of the SA-agarose resin of 373 nmol/mL, 75 μL of resin (Section 2.2.2, Step 2) will contain ~30 nmol of binding sites, which equates to a maximum concentration of 150 μM biotin in a typical 200 ul sample of MNase digested soluble yeast chromatin. Note that biotin is not removed in the steps after modification, before exposure of the soluble chromatin to the resin. Indeed, when the number of biotin molecules exceeds the streptavidin agarose binding sites, we observe that most of the spiked-in radiolabeled and BM modified H2B S112C- nucleosomes appear in the flow-through fraction, and only a small amount of labeled nucleosomes in the eluted fraction (Fig 4B, lanes 2 and 3). However, when the number of binding sites on the resin exceeds that of the biotin in the chromatin preparation, we observe highly efficient binding of the BM-modified radiolabled nucleosomes to the resin, with most of the nucleosomes appearing in the eluted fraction (Fig. 4B, lanes 4–7). These results indicate that concentrations of BM should be kept below about 150 μM in the modification step when using the resin as indicated. Also care should be taken to calculate the maximum biotin reagent that can be used for resins with other binding capacities.

We then sought to determine whether BM modified yeast nucleosomes generated by MNase digestion could be similarly specifically retained on the streptavidin-agarose resin. We carried out affinity purification of chromatin from BM-modified WT and H2B S116C nuclei, and prepared libraries for sequencing from the collected elutates from the streptavidin-agarose resin. Analysis of input and DNA libraries via gel electrophoresis revealed an obvious enrichment of mononucleosome length DNA recovered in the H2B S116C affinity purification as compared to the wild-type (Fig. 5). Following NGS, we quantified the fold enrichment of DNA recovered from H2B S116C compared to wild-type using the 1 ng of mouse nucleosome DNA spiked into every sample as a standard (Section 2.4.1). We typically find a 25–50- fold enrichment of DNA for H2B S116C indicating that we had selectively retained BM modified H2B S116C nucleosomes over unmodified wild-type nucleosomes (the enrichment is 58-fold for the example shown in Table 1). Length histograms of wild-type and H2B S116C libraries also suggests that we have selectively retained BM modified nucleosomes. Input libraries show similarly digested chromatin with the majority of DNA fragments being ~150 bp for both wild-type and H2B S116C. As expected, affinity purification results in a dramatic reduction in the amount of 150 bp DNA fragments for wild-type (Fig. 6). Conversely, we find that the relative amount of mononucleosome length fragments in the affinity purified H2B S116C library is commensurate with that of the input.

Fig. 5:

Analysis of input and affinity purified DNA libraries via gel electrophoresis. Following PCR amplification, DNA libraries were analyzed via gel electrophoresis (2% agarose, 0.5X TBE, 0.04% SDS) to assess quality of libraries and relative amounts of input and affinity purified DNAs. For H2B S116C site, there is a commensurate amount of input DNA but substantially more DNA present for affinity purified samples as compared to the wild-type.

Table 1.

Reads for each sample are aligned to reference genome for yeast (SacCer2) and mouse (mm10). The ratio of aligned yeast reads with respect to aligned mouse reads is then determined. In order to determine fold-enrichment, the AP:input ratio is calculated for wild-type and H2B S116C using the yeast:mouse aligned reads ratio. Lastly, the AP:input for H2B S116C is then divided by that of the wild-type to determine fold-enrichment.

Sample	Total Reads	Total yeast align	Total mouse align	Total align	% align	% mouse DNA	Yeast: Mouse
WT Input	4361189	3754810	100685	3855495	88.4	2.6	37.3
WT AP	5493527	2777102	2457627	5234729	95.3	46.9	1.13
H2B S116C Input	6206648	5616611	155073	5771684	92.9	2.7	36.2
H2B S116C AP	4297230	4167669	65691	4233360	98.5	1.6	63.4

Figure 6:

Length histograms of input and affinity purified DNA libraries for wild-type and H2B S116C samples. Following alignment to SaCer2 reference genome, the relative abundance of mapped DNA fragments from 0 to 500 bp is determined. For both wild-type and H2B S116C, the input DNAs have a major peak at ~150 bp (mononucleosomes). However, for the affinity purified samples, there is substantially more mononucleosome length DNA in the H2B S116C sample than that of the wild-type.

We next mapped reads from the input and affinity purified libraries to the yeast genome and generated nucleosome occupancy plots aligned to the transcription start sites for all annotated yeast genes. A comparison of nucleosome occupancy plots shows that nucleosome positions are virtually identical in wild-type and H2B S116C input samples (Fig. 7A). This result indicates that the H2B S116C cysteine substitution does not affect chromatin organization around promoters or interactions with ATP-dependent nucleosome remodeling complexes (Fig. 7A). Comparison of the nucleosome occupancy plots from the H2B S116C input and affinity purified samples also shows they are nearly identical to each other (Fig. 7b). In both plots, the characteristic decrease in DNA reads just upstream of the transcription start site is observed corresponding to the nucleosome depleted region (NDR). It is worth mentioning that the NDR is not entirely protein free. For example, a number of proteins involved in facilitating transcription reside in this region. Even still, this does not manifest in the nucleosome phasing map of the H2B S116C affinity purification, consistent with selective purification of DNA associated with BM-modified nucleosomes. Taken together, our data indicates that our method can be used to map nucleosomes genome-wide based on the accessibility of a specific histone amino-acid on the core histone octamer surface.

Figure 7:

Nucleosome phasing maps aligned to transcription start sites for all yeast genes. A. Nucleosome phasing maps derived from input samples for wild-type (blue) and H2B S116C (orange) nucleosomes. Nucleosome center positions were mapped to SaCer2 reference genome and then aligned with respect to transcription start site for all genes. The nucleosome depleted region (NDR) immediately upstream of the transcription start site, the gene body and transcription start site are indicated. B. Comparison of nucleosome phasing maps for H2B S116C input and affinity-purified samples.

The flexibility of this system lends itself to a number of interesting directions. The accessibility of any specific nucleosome surface can be explored without employing histone-specific antibodies which can vary between lots and efficiency in ChIP. Regardless of the residue probed in our method, the same chemistry can be used allowing for more direct comparison between distinct nucleosome surfaces. One area of debate involves the possible existence of subnucleosomes in which the H3-H3 surface is exposed on transcriptionally active sequences in yeast and higher-order eukaryotes [35–39]. However, earlier studies detecting such species involved the application of MNase digested chromatin to Hg-affinity resins, which may have promoted exposure of internal nucleosome surfaces [40, 41]. Other techniques infer the existence of subnucleosome species due to a reduction in symmetrical formaldehyde-mediated histone-DNA crosslinking for defined nucleosome positions [42]. Our method offers the advantage of directly probing the accessibility of an internal nucleosome surface in the context of native chromatin to better report on histone removal and/or distortion of the canonical nucleosome structure. In addition, chromatin has long been thought to condense and fold into highly structured domains. However, some recent in vivo studies have suggested that these higher order structures may be relatively random in their formation or may sparingly exist in yeast [15, 28, 43]. Located near the nucleosome acidic patch, the H2B S116C site is a good candidate to investigate this question as many models of nucleosome packing in higher order chromatin structure indicate con-facial packing of nucleosomes along the protein surface [43, 44]. In addition to using biotin-maleimide to probe accessible histone surfaces, the employment of a photo-inducible crosslinker, 4-azidophenacyl bromide (APB) can be used to identify discrete histone-DNA contacts in vivo. Previous studies in vitro have used this reagent to investigate histone-DNA contacts in oligonucleosome arrays as well as how these interactions are mediated by chromatin binding factors such as HMGN and H1 [45–47]. Performing these studies in vivo would allow for the identification of the necessary histone-DNA interactions which help govern chromatin structure.

While we have performed these proof-of-principle experiments in yeast cells, we envision this technique being implemented in a metazoan system. Previous work in Drosophila using engineered gene clusters investigated various post translational modification effects through mutation of residues of interest within the core histones [48]. Therefore, it should be feasible to introduce cysteine mutations into these constructs to explore the accessibility of histone surfaces in a more complex organism such as Drosophila.

• Most genome-wide methods utilize detection of DNA accessibility

• We introduce a novel method for probing chromatin accessibility via histone surfaces, validated in S. cerevisiae

• Cysteine substitutions were introduced into core histones and accessibility was assessed via reaction with the thiol specific reagent, biotin-maleimide

• Streptavidin affinity purification results in ~25-fold enrichment of nucleosomes containing the cysteine mutant histone over wild-type, and yields nucleosome phasing maps comparable to MNase-seq

• Our method can be used to monitor exposure of both external and internal histone surfaces in the nucleosome