Studying epigenetics using ChIP

​​​Chromatin immunoprecipitation (ChIP) is a powerful technique that allows researchers to examine the interactions between epigenetic regulators and DNA in their natural context. With ChIP, researchers can identify specific genes and sequences where a protein of interest binds, across the entire genome, providing critical clues to their regulatory functions and mechanisms. By dissecting the temporal and spatial dynamics of protein-DNA interactions, ChIP provides insights into core biological processes and disease pathology.

​ChIP is exceptionally versatile, with use in a broad scope of applications. From looking at sequence-specific protein binding to global regulatory processes, ChIP gives researchers the tools to integrate discoveries and paint a comprehensive picture of complex epigenetic regulatory systems.

ChIP has played a central role in elucidating gene regulation, transcriptional machinery, and chromatin structure. Here are some of the key proteins you can detect using ChIP:

• Transcription factors: By examining where, and when, specific transcription factors bind across the genome, researchers have identified specific binding sites and sequences, pinpointed downstream gene activation, and revealed genome-wide regulatory programs of transcription factors. Further investigation by ChIP and other methods has revealed these transcription factors to be master regulators behind disease pathology, where they orchestrate epigenetic dysregulation that results in cancers, autoimmune diseases, allergy, and many others. By identifying these master regulators and their downstream genetic programs, ChIP has provided novel targets for diagnostic and therapeutic strategies against a wide variety of diseases.

• Transcriptional machinery. ChIP studies examining the binding of RNA polymerase II, and other components of transcription, reveal promoter and enhancer sequences and novel mechanisms of transcriptional regulation.

• Chromatin Structure. ChIP studies were pivotal in the discovery and characterization of the histone code. By mapping the locations of specific histone modifications and comparing to known gene activation states, researchers have documented how acetylation or methylation on particular histone residues influence gene activation or silencing and higher order chromatin structure. These histone modification signatures can now be used to predict those aspects of epigenetic regulation at specific regions of the genome via ChIP. As new histone modifications and chromatin regulatory elements are discovered, ChIP continues to be an essential tool for revealing the functions of these elements, and complexities of their interactions, in genomic regulation.

Combining ChIP analysis of multiple proteins is a powerful way to build a complete picture of genomic regulation. This approach enables researchers to study how different types of proteins, and protein complexes, interact spatially and temporally at specific sites along a particular gene, or across the entire genome, to regulate gene transcription (Barski et al., 2007).

ChIP expands the scope and precision of epigenetic research

ChIP facilitates the analysis of epigenetic mechanisms on a variety of scales, with unrivalled precision. These are some of the things you can achieve using ChIP:

• Local epigenetic mechanism:

  • Map a protein of interest to a specific gene or genomic region of interest

  • Identify specific binding site sequences of a protein of interest

• Genome-wide epigenetic programming

  • Localize a protein of interest, such as a transcription factor, at all of its binding sites across the genome

  • Map proteins and chromatin characteristics across loci

  • Compare enrichment of a protein-protein modification (eg histone acetylation) at different loci under different conditions

• Dynamic epigenetic processes

  • Quantify a protein/protein modification at an inducible gene over a time course

  • Reveal essential mechanisms of epigenetic regulation and dysregulation involved in the biological process of interest by comparing ChIP results across different cellular states, conditions, and time points, researchers can reveal

  • Different tissues reveal epigenetic programs and genes responsible for differentiation and cell-type-specific functions and characteristics.

  • Different cell cycle states reveal epigenetic programs and genes responsible for cell proliferation and cell cycle control, with implications for developmental processes and cancer pathology.

  • Disease vs. healthy cells identify critical genes and programs that are dysregulated, to reveal underlying disease pathology and novel targets for diagnosis and treatment

  • Treatment vs. no treatment reveals whether certain treatments or conditions may be effective at correcting epigenetic dysregulation that underlies disease pathology

The ChIP procedure utilizes an antibody to immunoprecipitate a protein of interest, such as a transcription factor, along with its associated DNA. The associated DNA is then recovered and analyzed by PCR, microarray or sequencing to determine the genomic sequence and location where the protein was bound. The procedure can be broken down into five parts (Figure 4).

In some cases, cross-linking of DNA and proteins may be required to stabilize their interactions, particularly for proteins that interact as part of large protein complexes and do not directly contact DNA. Crosslinking fixes these molecular interactions and freezes them at a particular point in time and is termed cross-linking ChIP (X-ChIP). In contrast, native ChIP (N-ChIP) which is performed without prior crosslinking.

Crosslinking is generally performed with formaldehyde, which reversibly crosslinks protein to DNA, RNA, and other proteins. Other chemicals like cisplatin can be used to selectively crosslink only between DNA and protein. Dual crosslinking with additional reagents may be required to study interactions between DNA and particularly large protein/RNA complexes. UV crosslinking is irreversible and therefore not compatible with ChIP.

Chromatin must be fragmented into small pieces for efficient immunoprecipitation and precise mapping of the target antigen to the genome. The size of the DNA fragments will ultimately determine the resolution of genomic mapping, so it is important to optimize the fragmentation protocol.

• N-ChIP provides the highest resolution mapping, with enzymatic digestion generating fragments the size of a single nucleosome at 175 base pairs

• X-ChIP relies on sonication to generate ideal fragment sizes of 200–1000 base pairs

Steps 1 and 2 are incredibly important to optimize for each experiment to get the highest quality DNA possible for subsequent ChIP analysis.

The protein of interest and DNA fragment to which it is bound are then immunoprecipitated. The chromatin mixture is incubated with an antibody to the protein of interest, and either agarose or magnetic beads overnight at 4oC to form bead/antibody/protein/DNA complexes. The beads are then collected by either centrifugation, for Protein A, Protein G, or Protein A/G agarose beads, or by magnetic tube rack, for magnetic beads, to immunoprecipitate the antibody/protein/DNA complex. Non-specific binding is removed with subsequent washes.

The antibody/protein/DNA complex is eluted from the beads with SDS and heat. Crosslinking of protein and DNA must then be reversed with NaCl and heat for X-ChIP experiments. Any protein and RNA present are then degraded with proteinase K and RNase A, and the remaining DNA is purified with either phenolchloroform extraction or PCR purification kit.

The purified DNA is then analyzed by qPCR, hybrid array (ChIP-on-ChIP) or next-generation sequencing (ChIP-seq) to identify and quantify the sequences that have been immunoprecipitated. These sequences are mapped to the genome to identify the genes and regions where the protein of interest was bound.

The aim of cross-linking is to fix the antigen of interest to its chromatin binding site. Histones themselves generally do not require cross-linking, as they are already tightly associated with the DNA. Other DNA binding proteins that with weaker affinities for DNA or histones may require cross-linking to hold them in place.

• ChIP for histone modifications is unlikely to require cross-linking

• Non-histone proteins such as transcription factors and proteins contained in DNA binding complexes will most likely require cross-linking

• The further away from the DNA your interaction of interest lies, the less effective ChIP will be without cross-linking

While N-ChIP and X-ChIP both require chromatin fragmentation to make interactions accessible to antibodies, they require different fragmentation procedures utilizing micrococcal nuclease or sonication, respectively (Neill et al., 2003).

N-ChIP

For N-ChIP experiments, enzymatic digestion with micrococcal nuclease should sufficiently fragment your sample into single nucleosomes (monosomes containing ~175 base pairs of DNA).

• Purified monosomes are not suitable for analyzing interactions with certain chromatin binders, like transcription factors, which often bind inter-nucleosomal DNA. Sonication is recommended for these instances.

• Nucleosomes are dynamic and may rearrange during the enzymatic digestion. This may be a problem for mapping certain areas of the genome, and any changes should be monitored with suitable controls (see detection controls for quantitative PCR). X-ChIP should be performed as a control to assess any dynamic and unwanted changes in the absence of cross-linking.

• Enzymatic cleavage will not produce entirely random chromatin fragments. Micrococcal nuclease favors certain areas of genome sequence over others and will not digest DNA evenly or equally. Certain loci could be overrepresented, while others may be absent, potentially impacting the accuracy of the data.

• To get consistency in digestions, aliquot stock enzyme after purchase and run a new time course with a fresh aliquot every time you set up an experiment. Although enzyme quality may vary over time in storage, the risk of variation within chromatin preparations (degree of compaction, etc) is far higher; one chromatin sample should not be treated as being the same as all others before it.

X-ChIP

Formaldehyde cross-linking restricts access of enzymes such as micrococcal nuclease to their targets, making enzymatic digestion ineffective in X-ChIP experiments. Instead, sonication is used to generate random DNA fragments of 500–700 base pairs (2–3 nucleosomes).

• Avoid foaming, which decreases energy transfer within the solution and decreases sonication efficiency. Sonication may also be affected by cross-linking time, cell density, or cell type.

• While sonicated chromatin can be snap-frozen in liquid nitrogen and stored at -80°C for up to two months, avoid multiple cycles of freeze-thaw.

• Although sonication theoretically does not exhibit preferential cleavage of the genome, in practice this is rarely the case.

• DNA fragment sizes are typically larger, affecting the resolution of assay. Regardless, fragments up to 1.5 kb resolve well for most purposes in ChIP. Micrococcal nuclease digestion can improve resolution in combination with sonication and may be useful with gentle or incomplete cross-linked samples.

Regardless of which fragmentation method is chosen, it is important to always run a fragmentation time course to optimize fragment size when setting up an experiment.

Table 5: Advantages and disadvantages of N-ChIP vs X-ChIP.

Not all antibodies are appropriate for ChIP experiments, and many antibodies are not of ChIP-quality or validated for ChIP applications. Choosing the right ChIP-grade antibody is essential for the success of your ChIP experiment. If not commercially available, or if you would like to try an antibody that is not yet ChIP-tested:

• Antibodies approved for IP, IHC or ICC applications are good candidates. Similar to ChIP, these applications recognize the protein’s native conformation, in contrast to western blot antibodies, which may only recognize the denatured peptide form.

• Antibody specificity is a major concern and should be fully-characterized before application in ChIP experiments. For N-ChIP applications, use peptide competition in western blot. However, for X-ChIP applications, this method will not guarantee antibody function as cross-linking can dramatically alter epitopes. Instead, compare ChIP and western blot results using that antibody to confirm equivalent performance.

• Ideally, antibodies for ChIP should be affinity-purified; however, many laboratories use sera as their antibody source and then overcome background problems that may arise with stringent buffers.​

ChIP protocols and data analysis can be complex, so it is critical to include the right controls to ensure that the experiment worked as intended.

Sample controls

It is crucial to include an input sample control that has not been immunoprecipitated in all DNA recovery steps for comparison to pulldown sample results. It is ultimately this comparison that normalizes the data to provide interpretable results.

When immunoprecipitating for histone modifications, purified histone H3, and H1 can be used as positive controls for the quality of the histone preparation (histone H1 is commonly used for X-ChIP). Meanwhile, calf thymus histone preparation should be used as a positive control histone sample for checking antibody specificity in western blot.

Antibody controls

Various antibody controls are important to ensure that immunoprecipitation was successful and to rule out the possibility of contamination. Here are some examples of key controls: 

• Positive controls for active gene loci: H3K4me3 and H3K9ac

• Negative controls for silent gene loci: H3K9me3, H3K9me, and H3K27me3

• Negative control for a non-chromatin epitope: anti-GFP antibody

• Negative IP control: isotype IgG antibody control or beads only IP

Also, chromatin remodeling may move or remove histones at a particular locus (eg an active promoter). To confirm the preservation of nucleosomes at particular genomic loci, use a control antibody against a non-modified histone such as histone H3. When analyzing histone modifications, normalize to histone content with an anti-H3 antibody.

Quantitative PCR controls

If analyzing data by qPCR, additional controls are necessary to ensure the quality of data analysis. Certain areas of the genome will purify better than others, and some nucleosomes may rearrange during enzymatic fragmentation. As a result, it is important to generate PCR primers to several regions in the starting material, as well as the purified/ChIP material, as controls for spurious results. Generate starting material by lysing the starting cells and take a sample for simple PCR of control regions in parallel with ChIP.

Also, during the qPCR stages, it is essential to perform positive and negative control qPCR for genomic loci where you know the protein of interest should or should not be bound. It is also necessary to perform a non-template control qPCR as a negative control to ensure there is no contamination in the PCR.

ChIP protocols must be optimized at multiple stages to achieve the best results. Here are a few things that may need a little extra optimization to give you the best ChIP results.

Cross-linking (X-ChIP only)

Formaldehyde is recommended for reversible cross-linking. Formaldehyde is an efficient DNA-protein crosslinker but not an effective protein-protein crosslinker making it difficult to ChIP proteins that do not bind directly to DNA. Alternative cross-linkers may be useful for cross-linking over various intermolecular distances.

Cross-linking is a time-critical procedure and should generally only last a few minutes. Excessive cross-linking can create several issues, including a reduction in antigen availability and sonication efficiency. For example, epitopes may be masked or altered, reducing antibody binding to the antigen and ensuing extraction of material from your sample. 

• Always optimize cross-linking conditions with a time-course experiment (2–30 min crosslinking)

• Quench formaldehyde and terminate the cross-linking reaction with glycine

• Cross-links between proteins and DNA are disrupted by proteinase K, which cleaves peptide bonds adjacent to the carboxylic group of aliphatic and aromatic amino acids, to further aid DNA purification

Chromatin fragmentation

It is critical to optimize your chromatin input by fragmenting the chromatin to the appropriate size. Fragment sizes should be less than 1 kb, but ideally, 200-1000bp. The best resolution can be achieved with MNase digestion to single nucleosome level of 175 bp. Perform a time course of chromatin digestion over 2–30 minutes, purify DNA and run a gel alongside a DNA ladder to determine which conditions and timing achieve the optimal DNA fragment size (Figure 5). Different factors require optimization between N-ChIP and X-ChIP protocols.

Antibody concentration

It is important to titer the antibody to optimize the signal to noise ratio. Start with 3–5 µg of antibody for every 25–35 mg of pure monosomes. For quantitative ChIP, you may need to match the amount of chromatin with the same amount of antibody. ChIP typically requires a large amount of primary antibody (1-10 µg per ug of beads). As with many techniques, it is essential to optimize the amount of antibody at the beginning before you run your experiment.

Wash buffers. 

Determine the correct composition for appropriate stringency of wash steps, typically between 250–500 mM NaCl or LiCl. Higher concentrations of salt and detergent will give cleaner results. However, balance must be achieved between low background and detrimental effects on the target. If the buffer is too stringent, it will destroy specific antibody interactions, resulting in low signal. If the buffer is not stringent enough, non-specific interactions will remain, resulting in high background. NP-40 can be used as a detergent, while RIPA is commonly used for X-ChIP.

Standard ChIP workflows require a large number of cells. Approximately 106 to 107cells as starting material, below which the assay is hindered by high background binding, poor enrichment efficiencies, and loss of enriched library complexity. However, these large sample sizes can be difficult to obtain, specifically when examining precious sample types like transgenic mouse tissues or clinical samples. To adjust for lower sample inputs, a number of strategies can be applied.

1. Improving enrichment efficiency and minimizing sample loss

Several adjustments to the ChIP workflow can increase enrichment efficiency and minimize sample loss for low input samples (Mao et al., 2013 and Dirks et al., 2016).

• The quality and properties of the sample itself are important considerations. Specifically, in formalin-fixed paraffin-embedded (FFPE) samples, over-cross-linking can cause problems. Methods to extract soluble chromatin from FFPE samples may help (Cejas et al., 2016).

• The kinetics of the IP with low concentrations of antigen can be optimized by modifying variables like buffer pH, ionic strength, and time of incubation (Reverberi et al., 2007).

• Broad DNA fragment size distribution hinders analysis of low-input ChIP, which can be remedied by more limited sonication and/or MNase digestion for more uniform fragmentation (Gilfillian et al., 2012).

• While bacterial DNA is sometimes used as a blocking agent to reduce background for standard ChIP, it is not advised for low-input ChIP as it carries through the assay and confounds data analysis. Other blocking agents such as inert proteins or mRNA can reduce background binding in low-input ChIP, without contaminating the data.

• Miniaturization of the assay into microwell formats facilitates automation and increases the concentration of the antigen (target transcription factor) during the IP workflow – this avoids the “dilution effect” of low antigen concentrations that favor dissociation of the antibody-antigen complex and decreases the efficiency of ChIP.

• Maximize sample retention with single-tube assay formats and the use of magnetic bead purification rather than phenol-based extraction after each assay step.

• The immobilization of antibody and washes to remove non-antibody bound material is often overlooked. Standard protocols use Protein A/G, but alternatives like epitope-tagged proteins may run the risk of over-expression and introduction of artefacts (Xiong et al., 2017).

• Abcam’s high-sensitivity ChIP assay employs a unique chimeric protein to capture the antibody-bound protein-DNA complex, offering significant advantages.

• The capture protein is smaller than Protein A or G and is coated at high density on the surface of microtiter plate wells, providing a much higher number of IgG immobilization sites in a smaller area, which in turn ensures efficiency and concentration of eluted DNA.

• The chimeric capture protein shows superior stability across a wider range of pH and salt concentrations, which allows for higher stringency wash conditions.

• Successful ChIP starting with just 2 x 103 cells or 0.5 mg tissue

• Relative enrichment factors > 500x

• Fast and easy 5-hour protocol from cells/tissue to enriched DNA

• Microplate assay format for flexibility in sample throughput and automation (can be used in single-well, 8-well strip or 96-well plate format)

2. Readout and downstream data processing platforms

In addition to the assay itself, the choice and optimization of downstream processing (ie sequencing, array, or PCR) and bioinformatic analysis are also important.

• The detection platform impacts the assay’s sensitivity. ChIP-sequencing (ChIP-seq) is the gold standard platform for high sensitivity, with consistently lower noise than ChIP-on-chip (ENCODE. (n.d.). ENCODE Platform Comparison).

• The most common issues in low-input ChIP-seq are high numbers of unmappable reads, PCR duplicates, and poor library complexity. Therefore library preparation must be optimized for low input samples by optimization adapter ligation to avoid amplification-derived error and bias. Maximizing the efficiency of ChIP enrichment, as described above, can also help (Schmidl et al., 2015 and Bolduc et al., 2016).

• Bioinformatic workflows should be adapted to take into account likely process-derived biases in the data (Kiddler et al., 2011).​

Examining epigenetic mechanisms in specific tissues can reveal essential elements of tissue-specific genetic programming, development, and biological processes. ChIP can be a valuable tool for examining roles and mechanisms of tissue-specific transcription factors, gene activation and other aspects of epigenetic regulation. To perform ChIP from tissue samples requires specialized chromatin preparation protocols to ensure quality input material and reliable results.

The amount of tissue required will depend upon protein abundance, antibody affinity and the efficiency of cross-linking. The following protocol was optimized using 5–15 µg of chromatin for each ChIP assay, with 30 mg of liver tissue for each ChIP/antibody. Exact chromatin concentration should be determined for each tissue type before starting the X-ChIP assay.

Click here for our ChIP from tissue samples protocol.

Once pulled down DNA fragments have been immunoprecipitated and purified, they can be analyzed by several different methods.

qPCR

Utilizes gene or target-specific primers to amplify known target loci among pulldown DNA

Limitations: 

• Must know the genome sequence of target regions to design primers for readout

ChIP-on-Chip

Employs microarrays to examine the presence of many loci of interest, specific domains, etc. across the genome. Pulldown and control samples are amplified and labeled with complementary fluorescent probes. Samples are combined and hybridized to a microarray of interest. The ratio of fluorescent signals indicates enriched regions where proteins of interest have bound.

Limitations:

• Requires large cell numbers

• Not sensitive to repetitive elements

• A large number of arrays are necessary to cover the entire genome

• Susceptible to amplification bias after the ChIP procedure

• Lower resolution than ChIP-seq

ChIP-seq

Most commonly used method for genome-wide analysis with improved base-pair resolution and none of the limitations of ChIP-on-Chip. Pulldown DNA and control samples are amplified, followed by high throughput sequencing of the fragments, which are then aligned to the genome. Overlapping fragments form a peak, indicating where the protein of interest was bound to the genome.

Data should always be normalized for the amount of starting material to eliminate errors introduced by uneven sample quantities. To normalize data, take the final amplicon value and divide it by the amplicon value of input material. For histone modifications, the immunoprecipitated material is usually normalized to the input amount and the amount of the relevant immunoprecipitated histone. For example, ChIP with an H3K4me3 antibody will be expressed relative to the input amount and the amount of H3 immunoprecipitated. Measuring the amounts (and quality) of starting material is the key to interpreting your results effectively.

While ChIP-seq data analysis can be complex, it is arguably the most important part of the experiment. Robust data analysis is key to accurate and reliable results. A combination of online tools makes data interpretation accessible to bioinformatics specialists and wet lab biologists alike. There are many resources that demonstrate how to extract reliable results from ChIP-seq data, and how to interpret data sets for successful ChIP-seq analysis (Hurtado et al., 2010 and Yan et al., 2013).

For a step-by-step tutorial on ChIP-seq data analysis, watch our webinar here.

This webinar covers the following ChIP-seq data analysis workflow:

1. Quality control of sequencing reads using FastQC, with a side-by-side comparison of a successful and failed experiment.

2. How to align or map the reads for a reference genome using Galaxy/Bowtie.

3. How to perform peak calling to identify enriched regions for the protein of interest using Galaxy/macs.

4. Binding signal visualization with UCSC genome browser.

5. De novo motif discovery: how to find the A motif enriched within the binding regions with MEME-ChIP.

6. How to assign binding sites to genes and get enriched genome ontology terms to discover potential biological function with GREAT.

7. Heatmap generation for binding signal representation with seqMINER.