Epigenetic regulation of genomic DNA for gene expression is important in cellular differentiation and the development of an organism. Epigenetics also contributes to human diseases. On a molecular level, epigenetics involves modification of the histone proteins and DNA that make up chromatin. Protein families have been identified which mediate these modifications, and characterization of the specific binding interactions are important to define the specificity and affinity of the interaction. This characterization is also important in developing drugs which inhibit these chromatin-binding proteins. This white paper summarizes how Isothermal Titration Calorimetry (ITC) is used for characterization of proteins involved in epigenetic regulation.
Epigenetics is the study of heritable changes in gene expression caused by non-genetic mechanisms, without alterations in gene structure or DNA sequence. The epigenetic state of a cell evolves during the cellular differentiation and development of an organism, and epigenetic changes are linked to cellular reprogramming. Because epigenetic mechanisms may also be responsible for the integration of environmental responses at the cellular level, they potentially play an important role in the development of some diseases.
Figure 1. Simplified view of chromatin and epigenetics machinery. Reprinted with permission[55]
Epigenetic regulation of gene activity is complex and not yet fully understood, and involves transcription factors, growth factors and perhaps hormones. Epigenetic processes also involve the modification of chromatin (Figure 1). A histone octamer, composed of two copies of each of the histone proteins H2A, H2B, H3, and H4, is wrapped by a strand of 145-147 bp DNA, forming a nucleosome core. Multiple nucleosomes pack together to form chromatin.
Epigenetic events likely involve covalent modifications of histones, DNA and RNA, as well as chromatin remodeling, micro-RNA mediation, and other changes to chromatin structure. The flexible N-terminal tails of histones contain a range of site-specific post-translational modifications (PTMs) called "marks," including methylation and acetylation of lysine, methylation of arginine, and phosphorylation of serine, threonine and tyrosine residues. The PTMs on histone tails form specific patterns, and are added, read and removed by specific enzymes in a sequence- and modification-specific manner, resulting in additional marks.The existence and/or absence of marks on histone tails provides unique docking sites for specific binding and/or release of downstream effector proteins, resulting in diverse biological functions including transcription regulation, cell cycle control, differentiation and apoptosis.
In addition to histone tail modifications, DNA methylation at the 5 position of the cytosine base constitutes another common covalent epigenetic mark, which can contribute to gene silencing.
Protein families have been identified which mediate these modification events. The chemical modifications are carried out by “writers" including histone acetyl-transferases, histone methyltransfersases, histone kinases and DNA methyl-transferases. Over 80 different writer proteins have been identified. The marks serve as docking modules for “readers" including proteins that recognize specific methylated or acetyated lysine marks on histones. Over 300 different reader proteins have been identified. The marks are removed by “erasers" including histone deacetylases and histone demethylases, and about 50 erasers have been identified.
The molecular basis for each module to directly write, read, or erase marks is in the individual protein domain. These domains belong to protein families that have multiple members based on the sequence or domain similarity and they often exist as part of larger, multidomain proteins and multiprotein complexes. For each possible mark, there is a corresponding family of writer, eraser and reader proteins.
Early epigenetics research focused on single domain-single mark relationships, and has since moved into studying multiple domain-multiple mark interactions, and the complex relationships and signalling among the so-called "epigenetic network." This means there are hundreds of proteins involved in epigenetic regulation that are potential targets for drug discovery.
The modulation of epigenetic mechanisms is relevant for many diseases, including cancers, metabolic disorders, inflammation, central nervous system diseases, and viral infections.The enzymes that catalyze histone PTMs, including histone methyl-transferases, histone demethylases, histone acetyltransferases, and histone deacetylases, are considered druggable targets. Drugs that bind reader proteins like bromodomains are also being studied.
For more information on epigenetics and drug discovery involving epigenetic proteins, refer to recent reviews[2,3,7,11,13,48,50,55,59,62].
Since chromatin binding and modifying activities are often found in multi-protein complexes, it is a challenge to describe individual chromatin-binding domains. Chromatin-binding proteins often discriminate between histone PTMs and sequence contexts using subtle affinity differences that appear critical to their function, and these are difficult to differentiate in cell-based assays. To understand the function, activity, and specificity of chromatin-binding proteins, each interaction needs to be defined by a series of biochemical and biophysical assays.
Researchers can use appropriate enzymatic assays for activity measurements of many writers and erasers. These assays can use purified proteins or crude nuclear extracts. Since reader proteins have no enzymatic activity, researchers need to measure the binding interactions by other biochemical and biophysical assays.
Characterization of the specificity and binding affinity for chromatin-binding proteins requires looking at the purified module, with desired activity, interacting with its histone mark, such as a specific bromodomain reader protein recognizing and binding to a histone peptide mono-, di- or triacetylated at specific position(s). In vitro biochemical assays include ELISA, CHiP (chromatin immunoprecipitation[52], histone peptide bead-based pull-down assays[45], and gel-shift assays. Some assays can be done with nuclear extracts, while others require purified proteins.
As targets involved in writing, reading and erasing histone methylation are in drug discovery pipelines, high-throughput assays to support screening and compound profiling has become more sophisticated. Synthetic histone peptide surrogates (with appropriate marks) are used in high-throughput histone-binding screening assays such as AlphaScreen®[33,76], AlphaLISA®[83], and SPOT assays[48,51]. A common high-throughput assay in drug-protein screening is differential scanning fluorimetry (DSF) where drug binding is assessed by a shift in protein thermal stability.
These biochemical and high-throughput screening assays give reasonable qualitative information about the specificity and relative binding affinity by rank ordering. It is important to keep in mind that qualitative biochemical assays for histone-binding proteins present a number of potential pitfalls in terms of non-specific binding, artifacts, reliability and reproducibility, as well as potential false positive/false negative results. For example, AlphaScreen® assays have day-to-day variability issues[32].
The binding affinity of an individual chromatin-binding interaction is specifically defined by precise measurement of its dissociation constant, KD. The smaller the KD, the tighter the binding affinity. Comparison of KD values is an important component of the identification of binding, assessment of binding selectivity, and specificity for a family of related proteins (like bromodomains) binding to different histone marks. KD values are also useful in the design and optimization of potential inhibitors.
The biochemical and screening assays listed above do not provide precise KD; many rank-order strength of affinity based on a readout (change in color, or gel band position or density), or are simply a binding/no binding assessment. AlphaScreen® assays give IC50 values, which are related to but not the same as KD, while SPOT assays give a relative affinity based on the spot intensity.
It is critical that any qualitative screen be validated by at least one biophysical binding assay to measure KD, for example, surface plasmon resonance (SPR), fluorescence polarization, microscale thermophoresis (MST), or isothermal titration calorimetry (ITC). Biophysical binding assays are orthogonal to primary binding screens and provide quantitative KD measurements. These serve to validate the initial screening data, and also detect false positives, false negatives, and detection limits inherent to the original assay. The dissociation constants for chromatin interfaces are in the nanomolar (tight binding) to micromolar (weak binding) range, and biochemical assays are not able to differentiate affinities in the low or high ends of this range. Biophysical assays often require purified proteins for quantitative results. Biophysical KD assays can be performed with wild-type and mutant proteins, and histone peptides with different modifications to further characterize the selectivity and function.
Biophysical assays also provide more information beyond KD, such as binding kinetics, binding thermodynamics, and binding stoichiometry. These measurements are offered by orthogonal assays, and provide further insight into the specificity and selectivity of the interaction, the mechanism of binding, and the structure of the active binding domain. Further confirmation and characterization comes from X-ray crystallography, nuclear magnetic resonance (NMR) and other structural studies.
This whitepaper discusses the features of the ITC assay, and explains how ITC is used to measure the KD of binding events involving epigenetic proteins and chromatin. Several case studies are presented, which show how the binding affinity and other parameters provided by ITC are used to characterize specific interactions, and design inhibitors.
ITC measures the heat changes associated with binding events, determining the binding affinity (or KD, the dissociation constant) of any biomolecular interaction. Modern ITC instruments measure KD values in the millimolar (weak affinity) to single-digit nanomolar (tight affinity) range. Binding enthalpy, entropy and binding stoichiometry can also be measured by ITC. ITC is used to characterize binding interactions involving proteins, nucleic acids and lipids. There are recent review articles on the principles and applications of ITC[23,81].
ITC is established in life sciences as the “gold standard” method for the study of binding processes, since it is label-free, sensitive, versatile, and can be used with different materials in a variety of buffers. ITC is an essential biophysical assay, included in thousands of citations in peer-reviewed scientific journals.
Figure 2 represents how ITC works. One binding partner (the “ligand") is placed in the ITC injection syringe, which is titrated into the sample cell and mixed with the second binding partner (the “macromolecule”). The ligand and macromolecule can be proteins, nucleic acids, small molecule compounds, lipids, carbohydrates, metals, or any other substances which interact with each other. Both binding partners are in identical buffers. The matched reference cell is typically filled with water.
The ITC system is programmed to inject a specific volume of ligand into the macromolecule solution at timed intervals. The injection syringe also has a paddle in the bottom, so the mixture is stirred throughout the experiment.
Malvern Instruments has several ITC instruments available, including MicroCal PEAQ-ITC and MicroCal iTC200. These are "power compensation" ITCs, which measure the temperature difference between the reference and sample cells, and change the thermal power compensation to bring the temperature difference back to zero. The power compensation correlates directly with the binding heat.
In the center of Figure 2 is a representation of the raw output of an ITC experiment. In this example the binding heat is exothermic; that is, the negative deflection of the heat change represents heat given off upon binding of ligand to macromolecule. Each peak represents the heat change caused by a single injection of ligand. The first peaks are large because most of the ligand that is injected is binding to the macromolecule. As the macromolecule becomes saturated with ligand, with further injections less ligand will bind and therefore less heat is generated. At the end of the experiment, very little ligand is bound and the small amounts of heat generated are due to the heat of dilution.
When the raw ITC data are uploaded into the ITC data analysis software, an integration baseline is automatically generated and the area of each peak is calculated and plotted as a function of the molar ratio of ligand to macromolecule in the cell. This is shown on the right of Figure 2. The black dots represent individual peaks. The curved red line is the best fit to a binding algorithm from which we determine the enthalpy of the interaction, the stoichiometry (N-value: related to the midpoint of the curve) and the affinity (related to the concentration and the sigmoidicity of the data).
The KD is related to the Gibbs free energy (∆G) of the interaction by Equation 1:
∆G = RTlnKD (Equation 1)
Where R is the gas constant and T is the temperature of the reaction (in K). ∆G needs to be negative for a process to occur. The tighter the binding affinity, the smaller the KD, and ∆G becomes more negative.
∆G is also related to the binding enthalpy (∆H) and binding entropy (∆S) by Equation 2:
∆G = ∆H - T∆S (Equation 2)
Characterization of binding enthalpy and entropy provide insight into the binding mechanism, including involvement of hydrogen bonding, hydrophobic effects, and conformational changes which occur during binding. Both thermodynamics and affinity are used to describe the specificity and structural information related to the binding.
Figure 2. How ITC works. Left: Schematic of ITC cells and injection syringe. Center: Representative raw ITC data from titration experiment. Right: Representative binding isotherm from ITC experiment, fit to one set of sites binding model.
ITC is an established binding assay for proteins involved in epigenetic regulation. Isothermal titration calorimetry has been used to research many hundreds of epigenetic readers, writers and erasers over the past 2 decades, with the majority of these studies published after 2010.
Several early studies which describe the use of ITC to identify and characterize chromatin reader protein-histone interactions have been published[30,31,38].
There are several reasons why ITC is used as a primary binding assay as well as a secondary/validation assay after initial biochemical screening assays:
ITC accurately measures KD from millimolar to nanomolar values, the typical range of epigenetic protein binding.
ITC measures the heat change associated with binding and is a universal detection system. This means that native proteins can be used, and there is no need for any kind of label, tag, marker, or modification.
ITC is a true in-solution method, and no immobilization is required, eliminating potential stearic/orientation issues, as well as non-specific binding to beads or other materials.
ITC does not need any specific reagents or antibodies, and requires minimal assay development.
ITC can be performed in essentially any biological buffer, including those containing common additives like salts, reducing agents, detergents, cofactors, etc.
ITC does not require any fluorescence measurements, so is not susceptible to potential artifacts in fluorescence-based assays.
ITC has no molecular weight limits, and can be used with any size molecule to evaluate protein-small molecule, protein-protein, and protein-nucleic acid interactions, as well as those of large multi-protein complexes.
ITC results are data-rich - besides KD, ITC measures binding thermodynamics and stoichiometry, useful in structural biology, structure-activity relationships, and drug discovery.
Design and software improvements have resulted in modern instruments which are more sensitive and use less protein compared to early ITCs, making the technique easier to use and more cost-effective. Also, it is a non-destructive technique, making it possible to recover the precious protein after the experiment. ITC instruments are also available in automated formats to increase throughput.
For more information on use of ITC and characterization of histone-binding proteins, refer to Malecek & Ruthenberg (2012)[45].
ITC measures the binding between protein domains and the nucleic acid component of chromatin. A recent article studied mammalian 5-methylcytosine (5mC) oxidase Tet3, and characterized the full-length isoform containing an N-terminal CXXC domain (Tet3FL)[35]. This CXXC domain was identified as an epigenetic reader protein. As a preliminary binding screen, the authors performed gel mobility shift assays using oligonucleotides containing all known cytosine C5 modifications: cytosine (C), 5-methylcytosine (5mC), 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC), and 5-carboxylcytosine (5caC). To accurately quantify the binding affinities of Tet30s CXXC domain with the unmodified DNA and the 5caC-modified DNA, the authors used a MicroCal iTC200 system to directly measure binding of the mouse Tet3 CXXC domain to a 12-mer palindromic DNA duplex centered around an unmodified CpG dinucleotide or a 5caCpG dinucleotide (CcaCG DNA) on both strands. ITC analysis indicated that the mTet3-CXXC domain bound to the unmodified CCG DNA with a KD of 1.88 µM but bound to the CcaCG DNA with tighter affinity (KD of 0.56 µM).
ITC was also used in this study to measure binding of the mTet3-CXXC domain to a 12-mer CmCG, CfCG, or ChmCG DNA duplex that replaced 5caC with 5mC, 5fC, or 5hmC. Results showed that the mTet3-CXXC domain bound to CmCG and CfCG with a KD of 5.59 and 6.13 µM, respectively, and binding was even weaker to ChmCG. When 5caC on one of the DNA strands was replaced with unmodified C, resulting in a hemi-carboxylated DNA (hemi-CcaCG), they observed a slightly reduced binding affinity for mTet3-CXXC, with a KD of 1.57 µM. Replacement of the unmodified C preceding 5caC with A, T, or G also led to a reduced binding affinity. The ITC data confirmed that the Tet3 CXXC domain is a specific reader for 5caC DNA within the CcaCG sequence.
The authors chose several CcaCG-interacting mTet3-CXXC residues for mutagenesis and used ITC to measure their effect on the binding affinities of mTet3-CXXC toward both CCG and CcaCG DNA . The first observation was that mutation of H81 to alanine eliminated the binding of the CXXC domain to both CCG and CcaCG DNA (in agreement with previous structural observations). Secondly, K88A mutation led to >27-fold weaker binding toward CcaCG DNA but only a 3-fold weaker binding toward CCG. These observations reinforced the proposal that residue K88, through direct interaction with 5caC, plays a key role in specific recognition between mTet3-CXXC and CcaCG DNA. Thirdly, the mTet3-CXXC T80A and Q82A mutations both led to a small decrease (weaker) binding affinity toward CcaCG DNA, consistent with the structural observations that both residues contribute to binding to CcaCG DNA. Together, the mutational studies and ITC assays lend strong support to our structural observations for the mTet3-CXXC-CcaCG complex.
ITC was also used to provide insight into the recognition of 5-methylcytosine oxidation derivatives by the SUVH5 SRA domain[61]. The SET- and RING-associated domain (SRA) reads the epigenetic mark 5-methylcytosine (5mC) on genomic DNA. The authors performed ITC binding and structural studies to investigate the molecular basis of the recognition of the 5mC and other modifications, by an SRA domain of the SUVH5 protein (SUVH5 SRA). Using MicroCal VP-ITC, they demonstrated that the SRA domain bound to the hydroxymethylated CG (5hmCG) DNA duplex in a similar manner to methylated CG (5mCG). Also, the SUVH5 SRA domain exhibited weaker affinity towards carboxylated CG (5caCG) and formylated CG (5fCG). ITC was used to show that SUVH5 SRA recognized 5hmC in a similar manner to 5mC, but exhibited weaker affinity towards 5hmC oxidation derivatives.
There are several journal articles which cite the use of ITC to study the interaction of a histone acetyltransferase binding its substrate, such as acetyl-coenzyme A or coenzyme A[1], or histone methyltransferase binding to the cofactor S-adenosyl methionine[15,79].
A 2014 study[70] looked at the epigenetic writer histone H4K20 methyltransferase Suv4-20h2 in complex with its histone H4 peptide substrate and S-adenosyl methionine (SAM) cofactor. Analysis of the crystal structure showed that the Suv4-20h2 active site diverged from the canonical domain configuration and demonstrated a high degree of both substrate and product specificity. Performing methyltransferase and biochemical binding assays, the authors showed that that the Suv4-20 family of enzymes took a monomethylated H4K20 substrate and generated an exclusively dimethylated product. The authors looked at Suv4-20h paralogs and a Suv4-20h mutant, binding to SAM, using MicroCal iTC200, and observed similar KD values (11 to 17 µM). To quote the authors: "Given the narrow range of binding constants, it is unlikely that SAM binding represents a physiologically significant difference between the paralogs." The ITC data did show differences in binding enthalpy and entropy, which the authors did not explore further. The thermodynamics could provide additional insight into the binding mechanism of the Suv4-20h to SAM.
Most epigenetic protein domain interactions involve binding to specific histones, so synthetic histones are often incorporated in ITC experiments, as well as other binding assays. ITC has been used to study writers such as histone methyltransferase and DNA methyltransferase binding with histones[12,14,15,24,43,79,85]. ITC has also been used to measure the KD of several erasers binding to histones, such as histone deacetylase[10],histone lysine demethylase[54], Jumonji domain[27,28,78] and Sirtuin[80,84]. ITC measured KD for these interactions in the range of 30 nanomolar to about 1 millimolar. Note that several studies also looked at the interactions of different modules, as well as interactions with histones.
Many researchers use ITC to characterize reader protein binding to histone peptides. These include 14-3-3[77], chromodomain[30,34,63,71], PHD fingers[5,8,53,60], PWWP domain[74], Tudor domain[6,22,49,58,73,75], WD40[46], and YEATS domain[41].
Bromodomains are a major class of reader proteins, with over 60 human proteins containing at least one bromodomain (BRD) module. BRDs specifically recognize ε-N-lysine acetylation (Kac) motifs on histones. Several research articles incorporate ITC data to study BRD interactions[29,31,64].
Scientists with the Structural Genomics Consortium (SGC) recently undertook a large-scale structural analysis of the human bromodomain family[18]. They developed a platform of recombinant BRDs, and subcloned all human BRDs into bacterial expression systems. After they purified the different BRD proteins, they identified interaction sites for BRD with SPOT peptide arrays that covered all possible lysine acetylation (Kac) sites of the human histones. From SPOT, they identified 485 interactions of BRDs of a single Kac site on a histone peptide. Binding was confirmed by ITC experiments showing KD values between 3 and 730 µM. They also used ITC to show that BRD peptide recognition was dependent on patterns of multiple modifications, rather than on a single acetylation site. It should be noted that values for KD determined in solution by ITC with SPOT assay (using cellulose-immobilized histone peptides) did not always correlate, suggesting that peptides linked to cellulose supports used in this study did not allow quantification of binding affinities by SPOT. The authors also noted that most of the ITC binding studies between BRDs and histone pepides had a binding stoichiometry of 1, indicating that one BRD module bound one histone peptide. However for bromodomain BRD4(1) the stoichiometry was 0.5, suggesting that two of the BRD4(1) modules bound to one diacetylated histone peptide, and one BRD4(1) module bound to one monoacetylated histone peptide.
Most researchers cited in this whitepaper used ITC primarily to measure and validate KD, while others, like Filippakopoulos et al. (2012)[18], also used stoichiometry in their characterization studies. ITC is the only biophysical assay able to measure KD, stoichiometry, and binding enthalpy in a single experiment. ITC is also the only biophysical assay that directly measures the binding enthalpy ∆H. Binding enthalpy is related to hydrogen bonding and other non-covalent interactions. When comparing results from two binding experiments, if one has a more negative ∆H, this can indicate that more hydrogen bonds are involved in that protein-ligand complex. More researchers are incorporating thermodynamics along with binding for structure-activity relationship (SAR) analysis.
In another study[36], ITC, computational, and structural studies were used to characterize different reader proteins which specifically recognize methylated lysine residues on histone proteins. The authors used MicroCal Auto-iTC200 to obtain binding affinities and thermodynamics of histone H3K4me3 (positively-charged methylated lysine), H3C4me3 (neutral carba analog) and H3G4 (neutral with glycine, rather than lysine at position 4) peptides binding to five reader proteins that specifically recognize H3K4me3 (the PHD zinc fingers of JARID1A, BPTF, TAF3 and the Tudor domains of the Royal family of SGF29 and JMJD2A). The five reader pockets are different in the aromatic cage composition and architecture, enabling examination of the specific effect of individual components of the aromatic cage on binding differences. All ITC binding assays had a stoichiometry of 1. Comparative ITC experiments for the associations of H3K4me3 and H3C4me3 showed that:
The positively charged H3K4me3 bound 2- to 33-fold more strongly than the neutral H3C4me3 to 4 out of 5 reader proteins that contain Trp as part of the aromatic cage (JARID1A, TAF3, BPTF and JMJD2A (Figures 3 and 4)
Association of the Kme3 side chain with the aromatic cage was more favorable (more negative) in enthalpy than the association of the neutral Cme3 group to the same cage (Figure 4)
Association of the Kme3 side chain was less favorable (more positive) in entropy than the association of the Cme3 group to the same aromatic pocket (Figure 4).
ITC binding and thermodynamics data, along with computational and structural results, provided evidence for the presence of the favorable cation–pi interactions as shown by the enthalpy-driven association of the naturally-occurring Kme3 with the electron-rich aromatic cage of reader proteins.
Figure 3. ITC data for JARIDA1A PHD3 reader domain (top row) and TAF3 PHD reader domain (bottom row) binding to histones H3K4me3 (left), H3C4me3 (middle) and H3G4 (right). Top of each ITC titration is raw data, bottom is binding isotherm fitted to one sets of site model. KD values are inset. Structures of the histones are on the far right. Reprinted with permission[36]
In contrast to other readers that contain at least one Trp residue, H3K4me3 and H3C4me3 bound to the tandem Tudor domain of SGF29 with similar binding thermodynamics, indicating the lack (or at least only a minor contribution) of cation–pi interactions in the association of Kme3 by the Tyr/Phe-containing half-aromatic cage of SGF29 (Figure 4). This result agreed with previous observations that the strength of cation–pi interactions depends on the nature of the aromatic ring.
Figure 4. Thermodynamic data from ITC experiments showing the binding of five reader protein domains to H3K4me3 or H3C4me3. Blue: ∆G; green: ∆H; red: -T∆S. Negative values are “favorable” processes, positive values are “unfavorable.” Data adapted with permission[36]
This work provided experimental and theoretical evidence that reader proteins predominantly recognize trimethyllysine via a combination of favorable cation–pi interactions, and involve the release of water molecules due to the association with the trimethyllysine side chain. This information has implications in rational drug design by specifically targeting the aromatic cage of readers of trimethyllysine.
The potential modulation of epigenetic mechanisms by chromatin-binding proteins is relevant for many diseases including cancers, metabolic disorders, inflammation, central nervous system diseases, and viral infections.
Epigenetic writers, readers, and erasers can all be considered potential druggable targets. Modifications such as acetylation and methylation are well characterized, and a deeper understanding of the mechanisms of other modifications can lead to future drug discovery. There are already several small molecule drugs which inhibit epigenetic target proteins, and more still in the development pipeline, with some already in the clinic.
ITC is used during the drug discovery process for hit validation, lead optimization, and mechanism of action studies. KD, thermodynamics, and stoichiometry from ITC can be used to characterize drug-target binding and move chemical hits into leads[20,26,40].
Development of small molecule inhibitors against a specific epigenetic domain is challenging, especially for reader proteins like bromodomains, since the target protein shares structural domains with other proteins in the same epigenetic target family. Filippakopoulos et al. (2010)[19] described a cell-permeable small molecule (JQ1) which binds competitively to bromodomains. JQ1 was designed using known structure-activity relationship (SAR) studies. After cloning and expressing all human bromodomains, they screened for binding to JQ1 using differential scanning fluorimetry (DSF) thermal shift assay. Binding of (+)-JQ1 significantly increased the thermal stability of all bromodomains of the BET family, suggesting tight binding. No significant stability shifts were detected for bromodomains outside the BET family, suggesting that this ligand was highly selective.
Since the effectiveness and sensitivity of DSF can vary between different proteins, ITC was used to precisely determine binding constants. Enantiomerically pure (+)-JQ1 bound with a KD of about 50 nM and 90 nM to the first and second bromodomains of BRD4, respectively. Comparable binding was also measured by ITC, and the first bromodomains of BRDT and BRD2 had 3-fold weaker binding. Also, (+)-JQ1 showed no detectable binding in ITC to bromodomains that exhibited minimal temperature shifts in DSF. JQ1 high potency and specificity toward a subset of human bromodomains was validated by co-crystal structures with BRD4.
In vivo studies showed JQ1 displaced the BRD4 fusion oncoprotein from chromatin, prompting squamous differentiation and specific anti-proliferative effects in BRD4-dependent cell lines and xenograft models. These data established proof of concept for targeting protein-protein interactions of epigenetic readers and provided a new scaffold for the development of chemical probes for the bromodomain family.
Baud et al. (2014)[4] used a “bump and hole” approach for design of small molecule inhibitors of bromodomains. The authors hypothesized they could use a chemical genetic approach, based on engineered shape complementarity between the bromodomain and a known small-molecule inhibitor (in this case I-BET), and eventually generate a high-affinity bromodomain-ligand variant pair.
A conserved hydrophobic residue on the bromodomain was mutated to a smaller residue to generate a “hole” in the protein. The mutant was targeted with analogs of a known inhibitor I-BET bearing a sterically bulky “bump” that can be accommodated by the mutant protein. Wildtype (WT) bromodomains bound the bulky analog more weakly as a result of a steric clash between the bump and the naturally-occurring residue. The authors developed an ethyl derivative of the existing small-molecule inhibitor I-BET, and showed via ITC and DSF that it bound leucine/alanine mutant bromodomains with nanomolar affinity with up to 540-fold selectivity, relative to wild-type bromodomains. Binding enthalpy from ITC was also negative and favorable. Cell culture studies showed that blockade of the first bromodomain alone was sufficient to displace a specific BET protein, Brd4, from chromatin. Expansion of this approach could help identify the individual roles of single BET proteins in human physiology and disease.
Other studies have used ITC in conjunction with bromodomain inhibitor/probe screening assays such as DSF and AlphaScreen®[9,16,21,37,47,56,57,65,82].
ITC has also been used to measure KD of probes and inhibitors to other epigenetic protein families such as chromodomains[67,72], histone lysine methyltransferase[17,42,44], WD40[66], histone deacetylase[68,69] and MBT[25].
James et al. (2013)[32] looked at potential inhibitors of the histone methyl-lysine (Kme) reader protein L3MBTL3. Since lysine methylation is a key epigenetic mark, inhibition of methyl-lysine (Kme) binding proteins can improve understanding of these regulatory mechanisms and potentially validate Kme binding proteins as drug discovery targets. The authors started with inhibitor UNC1215, the first known potent and selective small molecule chemical probe of a methyl-lysine reader protein, L3MBTL3. In the article, they discussed the design, synthesis, and SAR studies that led to the discovery of a new inhibitor with improved selectivity, when compared to similar proteins.
In this study, affinities for small molecule compounds binding to L3MBTL3 were determined by an AlphaScreen® assay, and these binding trends were confirmed by an orthogonal time resolved fluorescence resonance energy transfer (TR-FRET) assay. The direct binding affinity of compounds of interest was also confirmed by use of MicroCal Auto-ITC200. The combination of AlphaScreen®, TR-FRET and ITC provided high confidence in interpreting the reported SAR trends.
Compounds 2 and 56 showed promise as more selective L3MBTL3 inhibitors, so the authors used ITC to further quantify the level of selectivity, comparing binding to two similar proteins (Table 1). Compound 1 was the original compound UNC1215. Compound 2 bound L3MBTL3 with a tighter affinity (smaller KD) compared to L3MBTL1, resulting in about 150-fold increased selectivity. Compound 56 was even more selective for L3MBTL3, with approximately 400-fold selectivity by ITC. In comparison, ITC data for compound 5 revealed that the compound was 16-fold selective for L3MBTL3 over L3MBTL1, while compound 1 was 78-fold selective for L3MBTL3 over L3MBTL1. Looking at the structural data, together with the ITC data, it suggests that it is possible to improve the selectivity of potent L3MBTL3 inhibitors by varying the amines that bind to both the first and second MBT domain in the dimeric binding mode.
| Compound 5 | Compound 1 | Compound 2 | Compound 56 |
L3MBTL3 binding by AlphaScreen® IC50 (µM) | 0.071 | 0.064 | 0.17
| 0.13 |
L3MBTL1 binding by AlphaScreen® IC50 (µM) | 2.9 | 2.3 | >10 | 9.6 |
L3MBTL3/L3MBTL1 selectivity by AlphaScreen® | 41 | 35.4 | >59 | 74 |
L3MBTL3 binding by ITC KD (µM) | 0.38 | 0.12 | 0.47 | 0.35 |
L3MBTL1 binding by ITC KD (µM) | 6.2 | 9.4 | 0.68 | 132 |
L3MBTL3/L3MBTL1 selectivity by ITC | 16 | 78 | 145 | 380 |
Table 1. Results of orthogonal analyses for binding affinities of small molecule compounds to L3MBTL3 and L3MBTL1. Data adapted with permission[32]
The identification of chromatin-binding proteins using biophysical screening, followed by quantitative KD values from ITC, is an established method to look at binding specificity and selectivity. With the appropriate series of histone peptides, and different wild-type and mutant proteins of the same epigenetic family, ITC can determine the contribution of individual residues or modifications to the total binding affinity for a more complex substrate. Incorporating binding, thermodynamic and stoichiometry data from ITC, structural studies like NMR and X-ray crystallography, and in vivo studies, will ultimately lead to a complete understanding of these complex interactions. Biochemical screening, ITC, SAR and in vivo assays can then be used in rational drug design to create inhibitors specific for readers, writers and eraser targets.
1. Albaugh, B.N., Kolonko, E.M., and Denu, J.M. Biochemistry 49, 6375-6385 (2010) doi:10.1021/bi100381y.
2. Andreoli, F., Barbosa, A. J. M., Parenti, M. D., and Del Rio, A. Curr. Pharma. Design 19, 578-613 (2013) doi:10.2174/1381612811306040578.
3. Arrowsmith, C. H., Bountra, C., Fish, P. V., et al. Nat. Reviews 11, 384-400 (2012) doi:10.1038/nrd3674.
4. Baud., M.G.J., Lin-Shiao, E., Cardote, T., et al. Science 346, 638-641 (2014) doi:10.1126/science.1249830.
5. Berardi, A., Quilici, G., Spiliotopoulos, D., et al. Nucl. Acids Res. 44, 3448-3463 (2016) doi:10.1093/nar/gkw103.
6. Botuyan, M.V., Lee, J., Ward, I.M, et al. Cell 127, 1361-1373 (2006) doi:10.1016/jcell.2006.10.043.
7. Campbell, R. M. and Tummino, P. J. J. Clin. Invest. 124, 64-69 (2014) doi:10.1172/JCI71605.
8. Chakravarty, S., Essel F, Lin, T., and Zeigler, Z. Biochemistry 54, 5766-5780 (2015) doi:10.1021/acs.biochem.5b00617.
9. Ciceri, P., Müller, S., O’Mahony, A., et al. Nat. Chem. Biol.10, 305-312 (2014) doi:10.1038/nchembio.1471.
10. Clark, M. D., Marcum, R., Graveline, R., et al. Proc. Natl. Acad. Sci. 112, E3669-E3678 (2015) doi:10.1073/pnas.1504021112.
11. Conley, J. Drug Discov. World 24-39 (2015) http://www.ddw-online.com/enabling-technologies/p303671-epigenetic-targets-:-on-the-verge-of-becoming-a-major-new-category-for-successful-drug-research-summer-15.html
12. Dharmaraja, V., Lee, J.H., Patel, A. et al. J. Biol. Chem. 287, 27275-27289 (2012) doi:10.1074/jbc.M112.364125.
13. Du, J. and Patel, D. J. Biochim. Biophys. Acta 1839, 719-727 (2014) doi:10.1016/j.bbagrm.2014.04.011.
14. Du, J., Zhong, X., Bernatavichute, Y. V., et al. Cell 151, 167-180 (2012) doi:10.1016/j.cell.2012.07.034.
15. Eram, M.S., Bustos, S.P., Lima-Fernandes, E., et al. J. Biol. Chem. 289, 12177-12188 (2014) doi:10.1074/jbc.M113.523183.
16. Federov, O., Castex, J., Tallant, C., et al. Sci. Adv. 1, e150072313 (2015) doi:10.1126/sciadv.1500723.
17. Ferguson, A. D., Larsen, N. A., Tina Howard, T., et al. Structure 19, 1262-1273 (2011) doi:10.1016/j.str.2011.06.011.
18. Filippakopoulos, P., Picaud,S., Mangos,M. et al. Cell 149, 214-231 (2012) doi:10.1016/j.cell.2012.02.013.
19. Filippakopoulos, P., Qi, J., Picaud, S., et al. Nature 468, 1067-1073 (2010) doi:10.1038/nature09504.
20. Freire, E. Drug Disc.Today 13, 869-874 (2008) doi:10.1016/j.drudis.2008.07.005.
21. Gacias, M., Gerona-Navarro, G., Plotnikov, A. N., et al. Chem. Biol. 21, 841-854 (2014) doi:10.1016/j.chembiol.2014.05.009.
22. Garske, A. L., Craciun, G., and Denu, J. M. Biochemistry 47, 8094-8102 (2008) doi:10.1021/bi800766k.
23. Ghai, R., Falconer, R. J., and Collins, B. M. J. Mol. Recog. 25, 32-52 (2012) doi:10.1002/jmr.1167.
24. He, C., Li, F., Zhang, J., Wu J., and Shi, Y. J. Biol. Chem. 288, 4692-4703 (2013) doi:10.1074/jbc.M112.426148.
25. Herold, J. M., Wigle, T. J., Norris, J. L., et al. J. Med. Chem. 54, 2504-2511 (2011) doi:10.1021/jm200045v.
26. Holdgate, G. Meth. Mol. Biol. 572, 101-133 (2010) doi:10.1007/978-1-60761-244-5_7,
27. Horton, J. R., Engstrom, A., Zoeller, E. L., et al. J. Biol. Chem. 291, 2631-2646 (2016) doi:10.1074/jbc.M115.698449.
28. Horton, J. R., Upadhyay, A. K., Qi, H. H. et al. Nat. Struct. Mol. Biol. 17, 38-43 (2010) doi:10.1038/nsmb.1753.
29. Itzen, F., Greifenberg, A., Bosken, C.A., and Geyer, M. Nucl. Acids Res. 42, 7577-7590 (2014) doi:10.1093/nar/gku449.
30. Jacobs, S.A. and Khorasanizadeh, S. Science 295, 2080-2083 (2002) doi:10.1126/science.1069473.
31. Jacobson, R.H., Ladurner, A.G., King, D.S., and Tjian, R. Science 288, 1422-1425 (2000) doi:10.1126/science.288.5470.1422.
32. James, L. I., Korboukh, V. K., Liubov Krichevsky, L. et al. J. Med. Chem. 56, 7358-7371 (2013) doi:10.1021/jm400919p.
33. Janzen, W. P., Wigle, T. J., Jin, J., and Frye, S. V. Drug Discov. Today Technol. 7, e59-e68 (2010) doi:10.1016/j.ddtec.2010.07.004.
34. Jin, J., Shi, J., Liu, B., et al. Plant Physiol. 168, 1275-1285 (2015) doi:10.1104/pp.114.255737.
35. Jin, S.G., Zhang, Z.M., Dunell, T.L., et al. Cell Reports 14, 493-505 (2016) doi:10.1016/j.celrep.2015.12.044.
36. Kamps, J.J.A.G, Huang, J., Poater, J., et al. Nat. Commun. 6, 8911 (2015) doi:10.1038/ncomms9911.
37. Kempen, H. J., Bellus, D., Fedorov, O., et al. Lipid Insights 6, 47-54 (2013) doi:10.4137/LPI.S13258.
38. Lachner, M., O'Carroll, D., Rea, S. et al. Nature 410, 116-120 (2001) doi:10.1038/35065132.
39. Ladbury, J. E. Biochem.Soc. Trans. 38, 888-893 (2010) doi:10.1042/BST0380888.
40. Ladbury, J. E., Klebe, G., and Freire, E. Nat. Reviews Drug Disc. 9, 23-27 (2010) doi:10.1038/nrd3054.
41. Li, Y., Wen, H., Xi, Y., et al. Cell 159, 558-571 (2014) doi:10.1016/j.cell.2014.09.049.
42. Liu, F., Chen, X., Allali-Hassani, A., et al. J. Med. Chem. 53, 5844-5857 (2010) doi:10.1021/jm100478y.
43. Liu, N., Zhang., Z., Wu, H., et al. Genes Devel. 29, 379-393 (2015) doi:10.1101/gad.254425.114.
44. Ma, A., Yu, W., Li, F., et al. J. Med. Chem. 57, 6822−6833 (2014) doi:10.1021/jm500871s.
45. Malecek, K., and Ruhenburg, A. Methods Enyzymol. 512, 187-220 (2012) doi:10.1016/B978-0-12-391940-3.00009-3.
46. Margueron, R., Justin, N., Ohno, K. et al. Nature 461,762-767 (2009) doi:10.1038/nature08398.
47. McLure, K. G., Gesner, E. M., Tsujikawa, L. et al. PLoS One 8, e83190 (2013) doi:10.1371/journal.pone.0083190.
48. Muller, M. M. and Muir, T. W. Chem. Rev. 115, 2296-2349 (2015) doi:10.1021/cr5003529.
49. Musselman, C.A., Avvakumov, N., Watanabe, R. et al, Nat. Struct. Mol. Biol. 19, 1266-1272 (2012) doi:10.1038/nsmb.2435.
50. Musselman, C. A., Lalonde, M. E., Cote, J. and Kutateladze, T. G. Nat. Struct. Mol. Biol. 19, 1218-1227 (2012) doi:10.1038/nsmb.2436.
51. Nady, N. Min, J., Kareta, M. S. et al. Trends Biochem. Sci. 33, 305-313 (2008) doi:10.1016/j.tibs.2008.04.014.
52. Nelson, J. D., Denisenko, O., and Bomsztyk, K. Nat. Prot. 1, 179-185 (2006) doi:10.1038/nprot.2006.27.
53. Org, T., Chignola, F., Hetenyi , C., et al. EMBO Rep. 9, 370-376 (2008) doi:10.1038/embor.2008.11.
54. Pack, L. R., Yamamoto, K. R., and Fujimori, D. G. J. Biol. Chem. 291, 6060-6070 (2016) doi:10.1074/jbc.M115.696864.
55. Pande, V. J. Med. Chem. 59, 1299-1307 (2016) doi:10.1021/acs.jmedchem.5b01507.
56. Picaud, S., Strocchia, M., Terracciano, S. et al. J. Med. Chem. 58, 2718-2736 (2015) doi:10.1021/jm501893k.
57. Picaud, S., Wells, C., Felletar, I., et al. Proc. Natl, Acad. Sci. 110, 19754-19759 (2013) doi:10.1073/pnas.1310658110.
58. Pieters, B. J. G. E., Meulenbroeks, E., Belle, R., and Mecinović, J. PLoS One 10, e0139205 (2015) doi:10.1371/journal.pone.0139205.
59. Popovic, R. and Licht, J. D. Cancer Discov. 2, 405-143 (2012) doi:10.1158/2159-8290.CD-12-0076.
60. Qiu,Y., Liu, L., Zhao, C., et al. Genes Devel. 26, 1376-1391 (2012) doi:10.1101/gad.188359.112.
61. Rajakumara, E., Nakarakanti, N.K., Nivya, M.A., and Satish, M. Sci. Rep. 6, 20161 (2016) doi:10.1038/srep20161.
62. Rothbart, S. B. and Strahl, B. D. Biochim. Biophys. Acta 1839, 627-643 (2014) doi:10.1016/j.bbagrm.2014.03.001.
63. Ruan, J., Ouyang, H., Amaya, M. F., et al. PLoS One 7, e35376 (2012) doi:10.1371/journal.pone.0035376.
64. Ruthenburg, A.J., Li, H., Milne,T.A., et al. Cell 145, 692-706 (2011) doi:10.1016/j.cell.2011.03.053.
65. Schulz, D., Mugnier, M.R., Paulsen E.M., et al. PLoS Biol. 13, e1002316 (2015) doi:10.1371/journal.pbio.1002316.
66. Senisterra, G., Wu, H., Allala-Hassani, A., et al. Biochem. J. 449, 151-159 (2013) doi:10.1042/BJ20121280.
67. Simhadri, C., Daze, K. D., Douglas, S. F. et al. J. Med. Chem. 57, 2874−2883 (2014) doi:10.1021/jm401487x.
68. Singh, R. K., Lall, N., Travis S. Leedahl, T. S., et al. Biochemistry 52, 8139-8148 (2013) doi:10.1021/bi400740x.
69. Singh, R. K., Suzuki, T., Mandal, T., et al. Biochemistry 53, 7445-7458 (2014) doi:10.1021/bi500711x.
70. Southall, S.M., Cronin, N. B., and Wilson, J.R. Nucl. Acids Res. 42, 661-671 (2014) doi:10.1093/nar/gkt776.
71. Stroud, H., Do, T., Du, J., et al. Nat. Struct. Mol. Biol. 21, 64-72 (2014) doi:10.1038/nsmb.2735.
72. Stuckey, J. I., Dickson, B. M., Cheng, N., et al. Nat. Chem Biol. 12, 180-187 (2016) doi:10.1038/nchembio.2007.
73. Su, X., Zhu, G., Ding, X., et al. Genes Devel. 28, 622-636 (2014) doi:10.1101/gad.233239.113
74. Tesina, P., Cermakova, K., Horejsı, M., et al. Nat. Commun. 6, 7968 (2015) doi:10.1038/ncomms8968.
75. Vermeulen, M., Eberl, H.C., Matarese, F., et al. Cell 142, 967-980 (2010) doi:10.1016/j.cell.2010.08.020.
76. Wagner, E. K., Albaugh, B. N., and Denu, J. M. Methods Enzymol. 512, 161-185 (2012) doi:10.1016/B978-0-12-391940-3.00008-1.
77. Walter, W., Clynes, D., Tang., Y., et al. Molec. Cell. Biol. 28, 2840-2849 (2008) doi:10.1128/MCB.01457-07.
78. Wang, C., Zhang, Q., Hang, T. et al. Acta Cryst. D71, 1955-1964 (2015) doi:10.1107/S1399004715012948.
79. Wu, H., Zeng, H., Dong, A., et al. PLoS One 8, e83737 (2013) doi:10.1371/journal.pone.0083737.
80. Yu, J., Haldar, M., Mallik, S., and Srivastava, D. K. PLoS One 11, e0152467 (2016) doi:10.1371/journal.pone.0152467.
81. Velazques-Campoy, A., Leavitt, S. A., and Freire, E. Meth. Mol. Biol. 1278, 183-204 (2015) doi:10.1007/978-1-4939-2425-7_11.
82. Zengerle, M., Chan, K. H., and Ciulli, A. ACS Chem. Biol. 10, 1770-1777 (2015) doi:10.1021/acschembio.5b00216
83. Zhan, Y., Kost-Alimova, M., Shi, X., et al. Epigenetics & Chromatin 8, 37-56 (2015) doi:10.1186/s13072-015-0026-4.
84. Zhao, K., Harshaw, R. H., Chai, X., and Marmorstein, R. Proc. Natl. Acad. Sci. 101, 8563-8568 (2004) doi:10.1073_pnas.0401057101.
85. Zhang, P., Lee, H., Brunzelle, J.S., and Couture, J.F. Nucl. Acids Res. 40, 4237-4246 (2012) doi:10.1093/nar/gkr1235.
