Researchers have made remarkable progress in the understanding of the importance of chromatin and nuclear organization in the biological processes involving our genetic material, DNA [2, 3]. Chromatin is in essence DNA molecules wrapped around octamers of histone proteins and compacted and condensed to a different degree. In its broader definition chromatin includes other, chromatin-associated proteins. Chromatin structure varies from the highly compacted form, heterochromatin, to less compacted forms, collectively called euchromatin. The structure is dynamic; changes in the chromatin structure can influence nuclear processes that involve DNA, such as transcription, replication, repair, and recombination [4-7].

Chromatin is a dynamic structure that organizes the genome and potentiates external and internal signals, ultimately determining the expression of genes in different cell types, at different developmental time points, and in response to stimuli [2, 8, 9]. These changes in the genome readout are brought about without any actual changes in the DNA sequence and are defined as 'epigenetic' changes. Epigenetic changes in chromatin can be brought about by the addition of unusual histone proteins (histone variants), alteration of chromatin structure by ATP-dependent chromatin remodeling, the addition of chemical flags to the histone tails (histone modifications) and the addition of methyl groups to the bases of DNA (DNA methylation).

A range of chromatin-associated proteins can write (via the recruitment of chromatin modifying activities) and read a particular chromatin structure and function as effectors. The underlying chromatin structure determines the association of transcription factors, DNA repair proteins, and others. Consequently, our studies of chromatin structure and its role in nuclear processes have been greatly aided by the analysis of the binding of various proteins to chromatin and by the determination of the location of epigenetic marks [10].

Determining the association of a given protein with chromatin can take the shape of a yes or no query - is the protein chromatin-associated under given experimental conditions - or query the association with a precise genomic location. Additionally, the position of a particular protein or histone modification can be determined on a genome-wide scale. Table 1 lists the common methods for detecting chromatin association of proteins.

[enlarge]

Figure 1. Protocol for biochemical fractionation of chromatin.

A simple, small-scale biochemical fractionation can analyze the association of a given protein with chromatin (Figure 1). The appropriate fraction can be probed by Western blotting for the presence or absence of a particular protein. The information obtained does not allow for the determination of a location for the protein, only whether it is associated with chromatin. The method is simple and does not require large quantities of material and is therefore quite useful in many settings where obtaining sufficient material might be a problem such as detecting transient association and association under specific conditions and has been especially valuable for query association of proteins during particular points of the cell cycle. First developed for yeast cells [11, 12], the method was later modified for mammalian cells and is generally applicable to any cell type [13]. It is important to note that in the final fractionation step the pellet (P3, Figure 1) contains insoluble proteins bound to chromatin and the nuclear matrix. Half of P1 is treated with 0.2 U of micrococcal nuclease, chromatin- but not nuclear matrix-associated proteins will be released into S3 upon this nuclease treatment. Proteins with known associations are used as controls.

ChIP assays are used to determine the location of chromatin-associated proteins and/or their post-translational modification status. They rely on using antibodies that specifically recognize the protein of interest or the modified protein (e.g., histone H3 Lys 9 methylation) to immunoprecipitate and analyze co-immunoprecipitated DNA. Early methods relied on using mild lysis conditions to preserve protein-DNA interactions but applied only to proteins that were in direct contact with DNA. The development of formaldehyde crosslinking methods [16] allowed such an analysis to be extended to virtually any protein associated with chromatin (Figure 2). Crosslinking and immunoprecipitation (CLIP) technologies for protein-RNA interaction studies have been comprehensively reviewed [17]. Specific ChIP protocols for plant tissue have been published as well [18]. Commercial kits, such as ChIP-IT® Express Kit and Sonication Shearing Kit from Active Motif [19, 20], are available to alleviate the labor and time involved. For example, Lee YR et al conducted ChIP experiments with Cell Signaling SimpleChIP Enzymatic Chromatin IP Kit (catalog number #9003) to confirm that endogenous MYC binds to the promoter region of WWP1 gene [21]. Huang H et al employed Auto iDeal ChIP-seq for Histones Kit on IP-Star Compact Automated System from Diagenode to conduct H3K36me3 ChIP-seq [22].

The isolation of protein-DNA complexes from cells using antibodies directed against a specific DNA-binding protein is highly dependent on the conditions used to extract and immunoprecipitate. A crucial issue is how to solubilize complexes from the nucleus under conditions that maintain protein-DNA interactions. Several methods have been used successfully, but it is of note that conditions must be adjusted depending on the protein-DNA complexes queried [23-25].

Inherently the methods involve hypotonic lysis of cells, isolation of nuclei and subsequent solubilization of chromatin with a nuclease (DNase I or micrococcal nuclease - MNase [26] ) under low salt conditions, followed by immunoprecipitation using antibodies recognizing the protein(s) of interest, for example, antibodies against modified histones [26]. Protein-DNA complexes are best eluted from the immune complexes with peptides wherever possible to reduce contamination with non-specific DNA, which can be released under more stringent conditions. Extracted DNA can be cloned for further analysis, sequenced or used to probe tiling arrays.

[enlarge]

Figure 2. A general outline for formaldehyde cross-linking and chromatin immunoprecipitation.

This has become a powerful method for the study of dynamic protein-DNA association in the context of chromatin [27]. There are several steps to the Chromatin Immunoprecipitation protocol using formaldehyde cross-linking as outlined in Figure 2. Formaldehyde cross-linking allows us to probe for protein-chromatin interactions that might not involve direct binding to DNA. This cross-linking method will generate protein-protein, protein-DNA, and protein-RNA links and is therefore suitable for analysis of different types of chromatin constituents, as well as for transient associations. It has also been effectively used to analyze the presence or absence of post-translational modification of chromatin. The method was initially developed by Varshavski and colleagues [16, 28] and refined by Paro [29] in Drosophila. Before it was extensively used in other systems, it was first applied to yeast cells, and the two most widely used variations were adopted by two yeast groups [30, 31]. The steps of the technique as outlined in Figure 2 are universal to all ChIP protocols, but the details vary somewhat depending on the experimental system or the preference of the research group.

Moreover, they often need to be optimized when one applies the method for the first time to an innovative system. Below we will discuss each step in some detail, possible modifications and different versions of the protocol. Table 2 summarizes the steps.

Some different reactive chemicals can be used for cross-linking; however, formaldehyde is widely used as it has several advantages [32]. Formaldehyde is soluble in water, and it is active over a wide range of conditions (buffer, temperature, etc.). Most importantly, it penetrates biological membranes easily and thus cross-linking can be done with intact cells - this reduces the chance of relocation of complexes during the preparation of cellular and nuclear extracts. Formaldehyde is added directly to either the media in which cells are grown or to whatever other material is used (e.g., Drosophila embryos, dissected ovaries, etc.) to a final concentration of 1%. The fixation depends on both time and temperature; slower fixation occurs at lower temperatures. Typical conditions are 15 min at room temperature with occasional stirring, but time can vary anywhere between 5 min and several hours. When a short fixation time is desired, fixation is stopped by the addition of glycine to 0.125M final concentration. The duration of formaldehyde treatment needs to be determined experimentally for a given starting material and protein of interest and has to be the best compromise in the recovery of soluble chromatin without inactivation of the immunological determinants of the protein of interest. For example, one study crosslinked 2 × 10⁶ cells in their culture dishes with 1% formaldehyde for 10 min at room temperature, quenched the reaction with glycine, and washed twice with ice-cold PBS (containing protease inhibitors) [33]. Vodnala SK et al fixed CD8+ T cells with formaldehyde for 7 minutes before quenching with glycine [19]. Saito T et al cross-linked wild-type and ATG7-knockout HepG2 cells with 1% formaldehyde for 10 min [34].

The epitopes can be masked, hidden and inaccessible upon prolonged fixation. For proteins that are difficult to detect, shorter times, lower temperature and an even lower final concentration of formaldehyde may alleviate the problem [35]. For example, He M et al crosslinked 2×10⁷ cells in 2 mM disuccinimidyl glutarate (Thermo Fisher) at room temperature for 30 min for BAP1 ChIP [36].

Lysing the cells and obtaining soluble chromatin are intimately linked. Depending on the material some additional steps might be required (e.g. yeast - mechanical breakage, Drosophila embryos - dechorionation) [30, 31, 37-39]. For cells grown in culture and for most other material this involves resuspension in a suitable Lysis Buffer and sonication. Cells treated with formaldehyde are complicated to lyse; hence the sonication step serves to achieve both more efficient lysis and to fractionate chromatin. As an alternative to sonication, cross-linked nuclei can be partially digested with micrococcal nuclease [40]. There are two major Lysis Buffer systems used - the SDS-Buffer system and the FA Buffer, described in [30] and [31]. Some protocols use slightly different Buffers, but these two are the most widely used, as they have been proven to work well in some systems. The SDS Buffer offers many advantages over other Buffers - most efficient lysis of cells in combination with sonication, less variation in the size distribution of chromatin fragments and less background. The buffer contains 1% SDS which is diluted to 0.1% before immunoprecipitation. Such conditions are suitable for most but not all antibodies. If it is experimentally determined (see below) that immunoprecipitation is not efficient in this Buffer, then the FA Lysis Buffer is an excellent alternative. For sonication, the most reproducible results are achieved with a BioRaptor, for example, a Bioruptor UCD-200 instrument from Diagenode [33, 41], with the size of chromatin fragments averaging ~300 bp [42]. Misonix sonicator has also been cited [14, 43]. The fragmentation of chromatin is critical because its extent will influence the resolution of mapping experiments. ScreenTape D5000 from Agilent has been used to quality-check the fragmentation [44]. A slightly different protocol for lysis of mammalian cell was developed by the Farnham group [45, 46]. For example, He M et al prepared nuclear extracts using the truChIP High Cell Chromatin Shearing Kit from Covaris and sheard chromatin to 200-700 bp with a Covaris E220 sonicator [36]. Nott A et al used a Covaris E220 sonicator in ChIP-seq [47]. Covaris M220 ultrasonicator [48] and Branson Sonifier 250 [49] have been used as well.

The critical parameter at this step is the choice of antibody. If the antibody has not been previously used/shown to work in ChIP assays, preliminary testing is advisable. Note that not all antibodies marketed by companies as "ChIP antibodies" work in ChIP. An antibody that works well in immunoprecipitation may not work at all in ChIP. Some antibodies do not work well in 0.1% SDS, and immunoprecipitation has to be carried out in FA Lysis Buffer conditions. Monoclonal antibodies and antibodies raised against a small synthetic peptide may or may not work well, as the epitope might be masked after cross-linking; on the other hand, they may have a higher specificity. In general, polyclonal antibodies or a mixture of two or more monoclonal antibodies is advisable.

Once it is experimentally determined that an antibody can immunoprecipitate under cross-linking conditions, it needs to be tested for specificity under those conditions. This can be done by including a negative control in the actual ChIP experiment (IP from RNAi treated cells, mutant cells, etc.). For example, Chopra S et al included an irrelevant isotype control antibody and sequences of the Pri-miR-21 promoter as negative controls for ChIP analysis of XBP1s in human monocyte–derived dendritic cells [41]. The inclusion of negative controls is especially true if a whole-genome association is to be determined. This step can be optimized through commercial kits, such as Chromatrap spin column ChIP kit from Porvair or Magna ChIP A/G Chromatin Immunoprecipitation Kit from MilliporeSigma [33].

Note for epigenetic modification studies:Antibodies that are used to recognize post-translational modifications of histones need to be carefully checked and validated, many of these cross-react and will produce false positives, see [50, 51]. Recent efforts to address these issues have resulted in the characterization of a large number of antibodies [51, 52]. Antibody Validation Databases [53] is an excellent resource for researchers, who are also encouraged to contribute their test results.

Reversal of cross-link is necessary to purify the DNA. It is achieved by heating the samples for 6 - 12 hours at 65ºC. After reversal of crosslink, samples are treated with RNase and proteinase K (or Pronase), followed by phenol: chloroform extraction of the DNA [19] or through a DNA purification kit such as Zymo ChIP DNA Clean & Concentrator kit [36]. Earlier methods [29] called for CsCl gradient purification of DNA. However, it was shown that two successive phenol: chloroform extractions are sufficient in both lower and higher eukaryotes [30, 31, 35, 45, 54].

[enlarge]

Figure 3. PCR analysis of co-precipitated DNA.

Analysis of co-precipitated DNA can take several different forms, as summarized in Table 3.

When only a few chromosomal locations are being examined, the presence/absence of a specific DNA sequence in an immunoprecipitation is determined by PCR using sequence-specific primers. This process is also called ChIP-PCR.

PCR using site-specific primers can indicate the presence of a protein or post-translational modification (PTM) at a specific location. Primers are designed to amplify fragments between 100 and 400 bp in length. In addition to the region of interest (primers P1/P2 in Figure 3), primers must be designed to amplify a region of certain lack of binding to serve as a negative control (primers P3/P4 and P5/P6 in Figure 3). Multiplex PCR reactions can be separated on acrylamide gels [29, 30, 54], but there are several constraints to regular PCR: 1) fragments must be amplified with similar efficiency; 2) PCR reaction must be stopped while in the linear range of amplification; 3) usually no more than 3 sets of primers in one reaction. Increasing amounts of input are analyzed to ensure that the PCR is in the exponential phase. Specific binding is represented by amplification of only P1/P2 (representing the region of interest) in the IP, while all regions are amplified the input sample as they are represented equally in the genome (see Figure 3, INPUT lane). Additional essential control is an IP with a non-specific antibody (see IP vs. C, Figure 3), or an IP with an extract from cells lacking the protein. The sequences of interest should be amplified only in the IP reaction, and not the control reaction (or minimally so). A much more convenient and quantitative method is to analyze the DNA by Real-time PCR. The amplification efficiency of the different fragments is less critical, and pieces of almost identical size can be amplified. Increasing amounts of input DNA are used to construct standard curves for every primer set, and IP results are represented as percent of input DNA. A useful measure of enrichment of a specific region is to represent it as a ration between that region and a non-specific region [42]. It is essential to include all controls, as false positives are easily obtained with sensitive PCR applications. Conversely, a negative ChIP result is only meaningful if there are other positive results.

When the binding of a protein or the localization of histone modification is studied in the context of the whole genome, there are two approaches – hybridization of co-precipitated DNA to microarrays (ChIP-chip) and Next-Generation sequencing (ChIP-seq) of co-precipitated DNA. The ChIP-seq data can also be confirmed by ChIP-PCR [19]. Unrelated cells can be mixed initially into the experimental samples to calibrate the final sequence counts (a process called calibrated cChIP-seq) [55]. For example, Rhodes JDP et al mixed 500,000 HEK293 cells into 50 million Control or SCC1^DEG mouse ESCs before fixation and normalized uniquely aligned mouse reads against human spike-in reads [56]. ChIP-seq data stored at NCBI GEO can be a great resource. For example, Y Shwartz et al analyzed the ChIP-seq data (GSE31239) of mouse hair follicle lineage [57] for adrenergic receptors in their investigation of hair follicle innervation and stem cell activity [58].

Genome-wide studies are aimed at determining the global localization and have been made possible by the sequencing of some genomes. There are several different platforms for Next-Gen sequencing, although they all offer the parallel sequencing of more than millions of DNA molecules in real time. In the Solexa technology of Illumina, both ends of DNA fragments to be sequenced are ligated to specific adapters, which allow them to be attached to a solid surface. Such a library can be prepared through commercial kits like NEB Ultra II DNA Library Prep Kit [48], KAPA HTP Library Preparation Kit [14], Swift Biosciences Accel-NGS 2S Plus DNA Library Kit [43], NuGEN Ovation Ultralow System [36] or NEBNext Ultra II DNA Library Prep kit from New England BioLabs with barcoded adapters (NextFlex) from Bioo Scientific [47]. Each molecule is then amplified – 1000 molecules per 1 um cluster and 40 million clusters per experiment. Each cluster is sequenced by synthesis. This technology allows for multiplexing ChIP experiments, DNA from different IPs can be bar-coded, and multiple experiments can be sequenced in one run [59]. Similar technology is the SOLiD platform formerly of ABI [60, 61]. It is currently offered by Life Technologies, which also has another platform – Ion Torrent. For exmaple, Boettcher S et al prepared sequencing libraries with using SMARTer® ThruPLEX® DNA-Seq Kit from Takara and obtained sequences with an Illumina NextSeq™ 550 Sequencing System to study the binding of wild-type and mutant p53 proteins to different regions of chromatin in K562-TP53 isogenic cells [44].

[enlarge]

Figure 4. CUT&RUN compared to crosslinked ChIP (X-ChIP) and native ChIP (from He & Bonasio [1] ). CP stands for the chromatin-binding protein of interest.

CUT&RUN (Cleavage Under Targets and Release Using Nuclease) is based on entirely different principles (see Figure 4). To perform CUT&RUN, cells or crude nuclei preparations are first bound to magnetic beads. The nuclei are then incubated with the primary antibody, the secondary antibody, and finally incubated with a fusion protein containing a protein A domain and MNase. A “light” treatment with calcium ions triggers the MNase to cleave the chromatin, releasing the protein-bound DNA into the supernatant, leaving the unbound portion of the genome in the nuclear fraction. The cleaved DNA can then be isolated based on solubility and extracted from the supernatant and used for downstream processing such as Southern Blot or sequencing. In CUT&Tag (Cleavage Under Targets and Tagmentation), the MNase in CUT&RUN is replaced with a hyperactive Tn5 transposase loaded with sequencing adapters, thus the whole process from live cells to sequencing-ready libraries is accomplished in a single tube on the benchtop or a microwell in a high-throughput pipeline in one day [62]. CUT&Tag can be accomplished on single cells [63].

CUT&RUN is carried out in intact cells, without crosslinking, sonication and strong detergents, all of which can be a source of problems in standard ChIP experiments. The solution changes can be easily performed, making the protocol highly efficient. Limitations of ChIP include variation caused by crosslinking time [64], and the observation that protein content and active transcription at a locus can produce artifacts in ChIP data [65-67]. Whilst native ChIP (ChIP performed without crosslinking) [68] can be carried out to avoid some of these problems, it requires stable interactions between the protein and DNA and often fails to detect transient events.

MNase introduces breaks in the DNA at the genomic site dictated by the primary antibody. Carrying out the MNase treatments briefly and on ice reduces the activity of the fusion protein. This prevents MNase from cleaving elsewhere in the genome, as well as stopping the diffusion of the released fragments.

The benefits of CUT&RUN compared to standard ChIP-seq protocols include low background. This is because the reaction is performed in situ. In contrast to ChIP methodologies which involve the sonication of the entire genome, fragments are only generated around the sites of interest with CUT&RUN. It avoids many of crosslinking-associated artifacts commonly observed with ChIP. Besides providing more reliable data, the reduced background allows the protocol to be more cost effective, as less sequencing is required to achieve the same depth. The protocol can be completed in one day, thus it is more time-efficient. Finally, its ease of performance means it is amenable to automation, increasing the potential for high throughput applications.

Researchers have demonstrated the success of this method in both yeast and human cell lines [69]. They also found the method to be superior at detecting rare binding events, or events that were difficult to analyze by ChIP-seq, such as those at AT-rich sites of the genome. The method is applicable to even compact regions of chromatin by accurately mapping H3K27me3 and is appropriate for use on low cell numbers.

One article used CUT&RUN to probe the chromatin landscape in early human embryos [70]. Taking advantage of CUT&RUN’s applicability to small cell numbers, the authors were able to map the epigenetic changes that occur from parental-to-zygotic transition, finding a dynamic reprogramming distinct from patterns observed in model organisms. Another report used CUT&RUN to demonstrate the importance of vitamin C in the epigenetic reprogramming of the female germline [71]. The authors showed that vitamin C was required for the complete DNA demethylation and for H3K9me2 loss that occured during embryogenesis. A lack of vitamin C during this stage led to reduced fertility in adult mice.

Other applications of CUT&RUN include investigating SOX2-chromatin interaction [72], macrophage differentiation from monocytes [48], probing the chromatin environment during embryonic heart development [73], and transcription factor reprogramming during human embryonic stem cell differentiation [74]. Work has also been carried out by groups aiming to enhance the usability of CUT&RUN, including analysis tools [75-79].

Novel techniques to analyze epigenetic modifications include both experimental methods, which are based on modified ChIP immunoprecipitation, fluorescence resonance energy transfer (FRET), mass spectrometry and liquid chromatography, and computational models. Concerning the advanced methods of ChIP immunoprecipitation, PIXUL integrated with Matrix ChIP has recently been introduced [80]. This method utilizes ultrasound transducers and is characterized by high speed and reproducibility. Also, quantitative methylation-specific PCR (qMSP) has been developed to evaluate cell-free circulating DNA methylation and used as a marker in the diagnosis and prediction of treatment outcomes [81].

To study the regulation of chromatin structure by epigenetic histone modifications, FRET-based biosensors have recently been generated [82]. The biosensors have been inserted into nucleosomes to detect histone molecules using live cell imaging. Another biosensor has been developed using surface plasmon resonance and poly-purine hairpin probes, which bind ds-DNA [83]. This method has allowed the identification of DNA modifications by direct interaction with ds-DNA and methyl-cytosine measurement. Also, a recent study has applied a highly multiplexed mass spectrometry analysis to identify chromatic modifications at the single-cell level [84]. The results of the study have revealed that non-heritable factors mainly induce aging-induced chromatin modifications. Also, an ultra high-performance liquid chromatography (UHPLC)-based technique has been effective in tracking DNA modifications, such as a single Dalton change in cytosine-to-uracil switch [85].

In addition to the experimental methods, several computational methods have been developed to anticipate variability of gene expression due to histone modification. In particular, a computational differential analysis of epigenetic modifications (DiffEM) has been applied to detect the modification regions along cell differentiation process [86]. As to computational modeling, Monte Carlo simulations based on published ChIP-seq signals have been used to describe the mechanism and consequences of heterochromatin protein 1 binding to the highly methylated H3K9me3 [87]. The reported model has identified the specific region of H3K9me3 is crucial to stimulate segregation into the heterochromatic stage. Also, a deep learning framework DeepDiff, which encodes the spatial architecture of input signals, has been effective for modeling of histone modifications [88]. In addition, existing ChIP-Seq data, such as those hosted at NCBI GEO, can be analyzed to obtain new insights. M Pradas-Juni et al, for example, evaluated earlier MAFG overexpression ChIP-seq data to identify the involvement of LincRNAs in hepatic glucose metabolism [89].

Dr. Kathryn McLaughlin added the section on CUT&RUN in October 2019.