Preparation, Biochemical Analysis, and Structure Determination of Methyllysine Readers

Abstract

In-depth in vitro characterization of methyllysine reader domains and their association with cognate methyllysine substrates is essential to better understand fundamental mechanisms of chromatin regulation and to design targeted therapeutics that disrupt these interactions. In this chapter, we summarize commonly used methods for preparation, biochemical characterization, and determination of structures of methyllysine reader domains. We provide a detailed protocol for the preparation of a GST-tagged methyllysine reader domain and for analysis of histone-binding activities using a combination of pull-down, tryptophan fluorescence, and NMR assays, and describe initial steps toward crystallization of the complexes.

1. INTRODUCTION

Methylation of lysine residues has emerged as one of the major post-translational modifications (PTMs) in histone proteins. The side chain of lysine undergoes methylation on its ε-amino group yielding mono-, di-, or trimethylated species. In physiological conditions, addition of a methyl group does not eliminate the positive charge on lysine; however, it alters the hydrophobic character, charge distribution, and size of the lysine side chain as well as its ability to act as a hydrogen bond donor. Methylation is a reversible mark, which is generated by lysine methyltransferases (KMTs), whereas lysine demethylases (KDMs) catalyze the removal of this PTM (Black, Van Rechem, & Whetstine, 2012; Del Rizzo & Trievel, 2014). Various histone lysine methylation patterns have been shown to characterize specific genomic regions and chromatin states and are linked to particular biological outcomes such as active gene transcription, repression, and DNA damage response (Bannister & Kouzarides, 2011; Ernst et al., 2011; Kouzarides, 2007; Maunakea, Chepelev, & Zhao, 2010; Zentner & Henikoff, 2013).

Methyllysine marks are recognized by a set of protein effector domains, or readers of these PTMs. The list of identified methyllysine readers is growing very fast and currently includes ADD (ATRX-DNMT3-DNMT3L), Ankyrin, BAH (bromo adjacent homology), chromo-barrel, chromodomain (CD), double chromodomain (DCD), HEAT, MBT (malignant brain tumor), PHD (plant homeodomain), PWWP, SAWADEE, tandem tudor domain (TTD), Tudor, WD40, and zf-CW (zinc finger CW) (reviewed in Musselman, Lalonde, Côté, & Kutateladze, 2012; Patel & Wang, 2013). Comprehensive mechanistic, biochemical, and structural studies reveal a conserved mode for the methyllysine readout, which involves caging of the methylated lysine in an aromatic pocket, typically formed by two to four aromatic residues. Complex formation is driven by cation–π interactions between the methylammonium group of the lysine and the aromatic side chains of the reader, as well as hydrophobic and van der Waals contacts, and release of high energy water molecules (Hughes, Wiggins, Khorasanizadeh, & Waters, 2007; Kamps et al., 2015).

Aberrant catalytic activities of KMTs and KDMs result in alterations in histone methylation patterns and genomic instability. Misregulated functions of these enzymes have been linked to a wide variety of human disorders and there are a growing number of examples of methyllysine readers being implicated in disease (Brookes & Shi, 2014; Chi, Allis, & Wang, 2010; Greer & Shi, 2012; Helin & Dhanak, 2013; Pedersen & Helin, 2010; Portela & Esteller, 2010; Wigle & Copeland, 2013). In-depth characterization of the methyllysine reader domains is vital for a better understanding of fundamental mechanisms of chromatin regulation and is essential in the design of targeted therapeutics. In this chapter we describe our laboratories’ overall strategy for preparation, biochemical analysis, and determination of structures of the methyllysine readers.

2. PREPARATION OF METHYLLYSINE READERS

Most methyllysine reader domains can be easily overexpressed as fusion proteins in Escherichia coli and purified for biochemical and structural analysis. However, care must be taken to ensure a stable and properly folded protein construct is chosen for monitoring the interaction with methyllysine substrate, and some optimization may be required to select the best expression and purification conditions to ensure high yield and stability. For an excellent in-depth review on optimization of expression in E. coli see Gräslund et al. (2008).

2.1. Construct Design

Boundaries of an initial construct can be determined through comparative analysis of the primary sequence homology and the predicted secondary structure, followed by assessment of the best cutoffs to promote stability and solubility. Subsequently the expression plasmid must be selected for the desired purification protocol and ultimate application.

• (a)Analyze primary sequence of the domain via a bioinformatics research tool of choice, such as SMART (smart.embl-heidelberg.de; Letunic, Doerks, & Bork, 2015) to determine approximate domain boundaries.
• (b)The exact construct boundaries should be chosen such that the N- and C-terminal residues do not truncate a consensus sequence, interrupt predicted secondary structure regions (determined using, eg, JPred; Drozdetskiy, Cole, Procter, & Barton, 2015), and are not hydrophobic.
• (c)Once the construct boundaries have been selected, the sequence can either be subcloned out of the full-length gene or custom ordered. If subcloning is chosen, it is possible that the resultant DNA construct harbors “rare” codons and a compensating expression cell line, such as BL21-CodonPlus or Rosetta2 should be used for expression purposes. Custom synthesis of genes can be beneficial as codons can be optimized for bacterial expression.
• (d)An inducible expression plasmid should be selected (the most common being isopropyl β-D-1-thiogalactopyranoside (IPTG) inducible) with the desired fusion tag and cleavage site. The most common tags used for the expression of methyllysine reader domains are the glutathione S-transferase (GST) tag, the hexa-Histidine (6 × His) tag, and the maltose-binding protein tag. Tags are typically positioned at the N-terminus, but can also be placed at the C-terminus. Common cleavage sites, for postexpression cleavage of the fusion tag, include sites for PreScission, thrombin, Factor Xa, or TEV proteases. Care should be taken to ensure that the cleavage site sequence of choice does not occur anywhere in the domain sequence, and that the residual residues after cleavage are acceptable.
• (e)The final DNA construct should be sequenced to confirm that the proper coding sequence has been achieved and is in frame in the expression plasmid.

2.2. Overexpression in E. coli

Once a proper construct is cloned into an expression plasmid it can be overexpressed in E. coli. The choice of growth media depends on the desired application and can also be optimized for the best yield. Rich media such as Luria broth (LB), 2XYT, or terrific broth (TB) will often produce high protein yields and are sufficient for most biochemical and biophysical experiments and X-ray crystallography (unless specific labeling is needed for phasing purposes). For NMR spectroscopy it is often necessary to overexpress proteins in minimal media for ¹⁵N and/or ¹³C (as well as ²H, if desired) isotope enrichment. E. coli cells can be grown directly in minimal media, or if desired in rich media to proper OD₆₀₀, harvested, and transferred to minimal media before induction (Marley, Lu, & Bracken, 2001). Rich broths are commercially available as premixed powders. M9 media can be made as follows:

Recipe for M9 media

• Autoclave the following in 1 L of water

• 12.8 g Na₂HPO₄·7H₂O

• 3.0 g KH₂PO₄

• 0.5 g NaCl

• 1.0 g ¹⁵N-NH₄Cl

• 2 mL (1 M) MgSO₄

• 200 μL (1 M) CaCl₂

Sterile filter and add the following to the cooled autoclaved media. 5 mL vitamin solution (1 tablet of daily vitamin dissolved in 50 mL water, spin down insoluble portion before filtering), or equivalent of commercially available vitamin solution.

25 mL of 20% (w/v) glucose (¹³C-glucose should be used if carbon labeling is desired).

Proper antibiotics.

Overexpression

• (a)Transform the expression plasmid into cells of choice following the recommended protocol and plate onto LB agar plates supplemented with the proper antibiotics.
• (b)Inoculate one, or better two, preculture tubes (10 mL media supplemented with proper antibiotics) with colonies from the plate and incubate in a shaker/incubator at 25–37°C overnight.
• (c)Inoculate 1 L of media supplemented with proper antibiotics with one preculture.
• (d)Grow cultures in a shaker/incubator at 37°C until OD_600~0.4–1.0.The choice of OD₆₀₀ value at which to induce should be optimized for each construct. (Save a sample of uninduced cells for SDS-PAGE analysis.)
• (e)Induce protein expression with, eg, 0.1–1.0 mM IPTG. (For metalbinding domains appropriate salts should also be added at induction. Example for Zn-fingers add 50–100 μM ZnCl₂, note that adding too much can be toxic to the cells.)
• (f)Incubate induced cultures in a shaker/incubator at 16–37°C for 4–18 h. The choice of induction temperature and time should be optimized for each construct. Generally, it is best to induce at lower temperatures for longer times when overexpressing in minimal media to increase solubility. (Save a sample of induced cells for SDS-PAGE analysis.)
• (g)Harvest cells by centrifugation at ~15,000 × g for 30 min. Pellets can be stored at −80°C for up to several months.

Troubleshooting

Problem	Potential Cause and Solution
Leaky expression seen in preinduced samples	Switch to a cell line that suppresses background expression such as pLysS or pLysE cells
No expression seen in postinduction sample	The full-length protein is either not expressing or is being degraded in the cell. (1) Rare codons in the gene can lead to severely truncated transcripts. Check for the presence of rare codons and if needed either switch to a compensatory cell line such as BL21-CodonPlus or Rosetta2 or codon optimize the gene. (2) The rate of translation or induction time needs to be altered. Optimize the concentration of IPTG, the induction temperature and length of induction and analyze yield by SDS-PAGE. (3) Expression is low in the chosen media. Switch media.

2.3. Purification

For fusion proteins the first step is affinity purification utilizing the fusion tag. Lysis and wash buffers can be varied considerably to include buffer of choice (eg, Tris, phosphate, HEPES, MOPS), monovalent ion of choice (eg, potassium chloride, sodium chloride) as well as additives such as divalent salts, Arg, glycerol, etc. Below is an example set of buffers that have worked well for purification of a number of PHD fingers and Tudor domains in our hands. These buffers should be optimized for each construct. Many methyllysine-binding domains have a low isoelectric point (pI) and thus a pH of 6.5–8 is appropriate; however, this should be assessed for each construct (using, eg, the ProtParam tool; Gasteiger et al., 2005). A buffer pH that is at least one unit from the pI of the construct is recommended. Note that the pI of the fusion construct should be considered when determining the pH for the lysis buffer.

2.3.1. Affinity-Tag Purification (Per Liter of Culture)

The following protocol is written for purification of a GST-fusion methyllysine reader domain with a PreScission protease cleavage site that is stable in Tris and potassium chloride, but can easily be modified for other constructs. Note that the needed resin, method of elution, and cleavage conditions will differ depending on the choice of tag and cleavage site. It is advisable to follow the resin manufacturer’s recommendation as a starting point.

Lysis buffer

• 20 mM Tris, pH 7.5

• 250 mM KCl

• 0.5% Triton X-100

• 3 mM DTT (freshly made)

• Protease inhibitors (freshly made)

• 1 mM EDTA (be cautious in using with metal-binding proteins)

High-salt wash buffer

• 20 mM Tris, pH 7.5

• 250–500 mM KCl

• 3 mM DTT

• 1 mM EDTA (be cautious in using with metal-binding proteins)

Low-salt wash buffer

• 20 mM Tris, pH 8

• 150 mM KCl

• 3 mM DTT

• 1 mM EDTA (be cautious in using with metal-binding proteins)

Note: If EDTA cannot be used due to interference with metal binding, it is best to autoclave water for making buffers to avoid protease degradation during purification.

• (a)Resuspend the cells into 50 mL of lysis buffer by running through a 10 mL serological pipette. After resuspension lysozyme can be added if desired at 1 mg/mL and incubated for 30 min to increase lysis efficiency.
• (b)Lyse the cells by freeze/thaw, sonication, or homogenization methods.
• (c)Clear the lysate by centrifuging at ~25,000 × g for 30 min. (Save samples of the pellet and supernatant for SDS-PAGE analysis.)
• (d)Incubate the supernatant with glutathione agarose or glutathione sepharose resin for at least 1 h rocking at 4°C. (Lysate can also be loaded onto a prepacked glutathione column using a fast protein liquid chromatography (FPLC) pump.)
• (e)Wash the resin with 100–500 mL of high-salt wash buffer (for some domains a high-salt wash is unnecessary) followed by 100–500 mL of low-salt wash buffer. Every 100 mL of flow-through should be analyzed either by UV absorbance or SDS-PAGE to determine the appropriate wash volume. Once no protein is detected in the flow-through washing is complete.
• (f)If the cleaved domain is desired, skip to step (g). If the GST-fusion domain is desired, incubate the washed resin with ~20 mL of wash buffer supplemented with 50 mM L-glutathione (note that L-glutathione is very acidic, therefore make sure to pH the stock solution). Elute the GST-fusion domain from the column and wash the resin with an additional 50 mL of buffer to collect any residual protein. Be sure to remove glutathione by dialysis or concentration/exchange before further application.
• (g)If cleaved protein is desired, incubate washed resin with ~20 mL of low-salt wash buffer supplemented with PreScission Protease, rocking overnight at 4°C. Then elute cleaved protein and wash resin with an additional 50 mL of wash buffer to collect any residual protein.
• (h)Analyze all samples by SDS-PAGE (Fig. 1). (Antibodies to fusion tags are commercially available and can also be used for analysis.)

Fig. 1.

SDS-PAGE analysis of expression and purification steps for a GST-fusion methyllysine reader domain, stained with Coomassie Blue. Shown are (1) ladder, (2) cells preinduction, (3) cells postinduction, (4) lysate pellet, (5) lysate supernatant, (6) flowthrough from glutathione column, (7) wash #1, (8) wash #2, (9) wash #3, (10) wash #4, (11) glutathione beads prior to PreScission protease cleavage, (12) elution, (13) glutathione beads after cleavage with PreScission protease, and (14) concentrated elution. For reference, the fusion domain is boxed in lanes 3 and 11 and the cleaved methyllysine reader domain is boxed in 14.

Troubleshooting

Problem	Potential Cause and Solution
The protein expresses but is only in the pellet after lysis	It is wise to first confirm expression and presence of the fusion protein in the pellet through Western analysis. If confirmed, this suggests that the protein is insoluble. There are several potential causes and solutions to this. (1) The lysis buffer could be incompatible with the protein, check the composition of the buffer especially the pH and salt concentration and alter to increase solubility. (2) The cells were not lysed, try relysing, another lysis method or a combination oflysis methods to increase lysis efficiency. (3) The protein misfolds during expression. This can usually be addressed by slowing down the rate of expression. Lower the concentration of IPTG for induction and/or lower induction temperature
The protein does not bind to the resin	The potential cause ofthis depends on the tag. (1) Confirm that the binding buffer is compatible with the resin. (2) If using a structured fusion tag, it may indicate that the tag unfolds, which could occur either during expression or lysis. Try different induction conditions described above for solubility. Or try altering the lysis method to a less harsh method. (3) The resinbinding site on the tag may be occluded by the domain ofinterest. Try altering the linker between the fusion tag and domain, switching termini for placement of the tag, or switching tags
There is no cleavage reaction	The cleavage site may be occluded. (1) Try eluting off the resin first and cleaving in solution. (2) Try altering the linker between the fusion tag and protein, switching the termini that the tag is on, or switching tags
The protein cleaves but the tag-free domain does not elute from the resin	It is likely that the cleaved domain is either sticking to the resin or precipitating after cleavage. (1) Try eluting with a higher salt buffer to reduce sticking. (2) Try cleaving after eluting the fusion protein from the resin. (3) Alter cleavage conditions to promote solubility ofthe tag-free domain. (4) Switch tags

2.3.2. FPLC Purification

For some biochemical assays such as pull-downs, the affinity-tag purification step may yield a sufficiently pure sample. For most biophysical assays and structural determination it is best practice to incorporate further purification steps, for example, utilizing FPLC (Fig. 2). A common procedure would incorporate cation or anion exchange (eg, a Source S or Q column purification) and/or size-exclusion chromatography (eg, Superdex 75 or Sephacryl S-100 column purification) in order to yield high-purity samples.

Fig. 2.

FPLC purification of a methyllysine reader domain. (A) Chromatogram from Source 15Q purification. The UV trace at 280 nm is shown in black and the percentage of 1 M KCl buffer is shown in gray. SDS-PAGE analysis of the major peak is shown in the inset. (B) Chromatogram from the subsequent Superdex 75 purification. The UV trace at 280 nm is shown and the SDS-PAGE analysis of the major peak is shown in the inset.

• (a)Concentrate the elution from the affinity resin to 0.5–2.0 mL (depending on the size of the FPLC loading loop). Note that for a cation/anion exchange column a super loop can also be used to load the elution.
• (b)Spin down the sample at 5000 × g for 3 min to remove any insoluble particles and load onto the appropriate column that has been preequilibrated into the desired buffer. Run according the manufacturer’s recommendation.
• (c)Check the purity of fractions by SDS-PAGE, pool the clean fractions and concentrate the sample to the desired concentration for further purification or application.
• (d)At this point many methyllysine reader domains are compatible with long-term storage at −80°C. It is recommended to flash freeze the protein before storage. Protein stability should be confirmed by SDS-PAGE after thawing to ensure the freeze thaw process has not damaged the domain.

Troubleshooting

Problem	Potential Cause and Solution
Protein does not bind the cation/anion exchange column	(1) The wrong ion exchange column is being used. Assess the pI (and pH of the loading buffer) and confirm that the domain of interest will bind to the chosen column. (2) The salt concentration in the loading buffer is too high. Exchange into a lower salt buffer before loading
The protein has the incorrect retention volume on gel filtration based on expected for the molecular weight	Run standards in the chosen buffer to confirm column performance and where the protein should be eluting. (1) The protein may have degraded. Check by SDS-PAGE and alter conditions to reduce degradation. (2) The protein may be unfolded. This can be confirmed by another method such as circular dichroism. Construct boundaries may need to be altered to obtain a well-folded domain, or expression/ purification conditions may need to be altered to promote folding. (3) The protein is aggregating. Alter buffer composition to decrease aggregation

3. SCREENING FOR METHYLLYSINE RECOGNITION

A simple way to screen for methyllysine-binding activity is through the pull-down assay. The following protocol assumes a GST-fusion methyllysine reader domain that is stable in Tris and potassium chloride, and biotinylated methylated peptides. Other combinations of fusion tags can also be used. There are a large number of biotinylated methylated peptides commercially available. As lysine can be mono-, di-, or trimethylated it is wise to obtain all possible methylation states of the substrate of interest as well as the unmodified counterpart for screening.

Example binding buffer

• 20 mM Tris, pH 8

• 150 mM KCl

• 0.5% Triton X-100

• 3 mM DTT (freshly made)

Example wash buffer

• 20 mM Tris, pH 8

• 150–250 mM KCl

• 0.5% Triton X-100

• 3 mM DTT (freshly made)

• (a)Prepare protein and peptide stocks at 1 μg/μL.
• (b)Mix 1 μg of GST-fusion domain with 1 μg of biotinylated peptide and bring to 100 μL with binding buffer. Also prepare a negative control of GST-fusion alone. Incubate overnight at 4°C rotating end-over-end.
• (c)Add 25 μL of streptavidin sepharose or streptavidin agarose slurry (~50% slurry equilibrated in binding buffer) to each sample and incubate for 1 h at 4°C rotating end-over-end.
• (d)Centrifuge for 2 min at 2500 × g and remove supernatant.
• (e)Add 100 μL of wash buffer, invert 5–10 times and centrifuge for 2 min at 2500 × g. Repeat this wash step three more times.
• (f)After the final wash add 25 μL of SDS loading dye to the resin and boil samples for 10 min. Also prepare a 10% input (0.1 μg fusion protein in 25 μL loading dye).
• (g)Run samples (including negative and positive controls) on an acrylamide gel.
• (h)Analyze by Western using an anti-GST antibody (Fig. 3).

Fig. 3.

Pull-down experiment with a GST-fusion methyllysine reader domain using a series of biotinylated histone peptides each containing a single modification as denoted, detected using an anti-GST antibody.

Troubleshooting

Problem	Potential Cause and Solution
Protein is detected in the negative control	The GST-tagged domain is sticking to the resin. Try increasing the salt concentration of the washes, Triton X-100 concentration and/or the number of washes
Signal is very weak as compared to positive control	The wash conditions may be too stringent. Try lowering the salt concentration or Triton X-100 concentration in the wash buffer

4. MEASURING AFFINITY FOR METHYLATED SUBSTRATE

There are many methods appropriate for determining the affinity of methyllysine reader domains for substrate peptides, including isothermal titration calorimetry (ITC), surface plasmon resonance (SPR), and fluorescence anisotropy. Often there is a tryptophan in the aromatic cage or otherwise in the binding pocket of these domains and a dearth of tryptophans in the peptide, making intrinsic tryptophan fluorescence spectroscopy an excellent method as well. Below is a procedure for affinity determination by tryptophan fluorescence spectroscopy (Fig. 4). (Note: intrinsic tyrosine or phenylalanine fluorescence can also be used, though the excitation and emission wavelengths should be altered accordingly.)

Fig. 4.

Monitoring peptide binding of a methyllysine reader domain by intrinsic tryptophan fluorescence. (A) Corrected fluorescence spectra in the presence of increasing concentration of peptide. (B) Nonlinear regression fit of the change in fluorescence from apo as a function of peptide concentration.

• (a)Freshly prepared protein should be extensively dialyzed into a freshly made appropriate binding buffer and peptide stocks should be prepared in the same buffer.
• (b)A quartz cuvette should be extensively washed and dried, then filled with buffer.
• (c)The excitation wavelength should be set to 295 nm and emission spectra recorded from 305–405 nm with a 0.5 nm step size and a 1 s integration time. For best results collect and average three scans.
• (d)Replace buffer with the protein sample and collect another emission spectrum. The appropriate protein concentration should be determined based on the signal to noise ratio, keeping in mind that fluorescence intensity will likely decrease significantly at saturation. For domains containing a single Trp, concentrations of 0.5–10 μM generally yield good signal for the suggested data collection parameters, however this may take some optimization.
• (e)Add progressively increasing concentrations of peptide in a step-wise manner, keeping the total volume increase as small as possible while maintaining good pipetting accuracy. If equipped, setting a low stir speed is helpful to obtain proper mixing of peptide, otherwise the peptide should be mixed by pipetting. Upon each addition of peptide collect another emission spectrum.
• (f)Reference titrations should be performed adding buffer to protein alone as well as adding peptide to buffer alone.
• (g)Subtract the buffer scan from each sample run and make any other necessary corrections such as for dilution, photobleaching, and inner filter effects (see Kozlov, Galletto, & Lohman, 2012 for details on such corrections). Determine the maximum fluorescence signal for each corrected data set.
• (h)Compute the change in fluorescence intensity as compared to apo protein for each titration point and determine the K_d by nonlinear least-squares analysis (Fig. 4).

Troubleshooting

Problem	Potential Cause and Solution
A significant decrease in fluorescence is detected in the buffer reference titration	This can occur for several reasons. (1) The protein may be photo bleaching. Increase the concentration of protein to obtain sufficient signal throughout the titration and adjust for photobleaching in analysis. (2) The protein may be aggregating. Adjust conditions to reduce aggregation. (3) The protein may be sticking to the pipette tip, or to the cuvette. To assess if it is sticking to the cuvette collect several spectra at different concentrations and check for linear dependence of signal. If protein is sticking to the pipette obtain low- retention tips. Pipette tips can also be treated in the lab using a siliconizing agent
An increase in fluorescence is detected in the buffer reference titration	The buffer is likely contaminated. Fresh buffer should be used for all fluorescence experiments
An increase in fluorescence is detected in the peptide reference titration	Confirm that the substrate is not intrinsically fluorescent. If not, the peptide stock may be contaminated

5. ANALYSIS OF BINDING USING NMR SPECTROSCOPY

Nuclear magnetic resonance (NMR) spectroscopy is a very powerful tool for analysis of complex formation. NMR can be used to map binding pockets, determine dissociation constants, examine conformational dynamics on a wide range of timescales, and determine atomic resolution structures. Here we outline procedures for sample preparation and analysis of chemical shift perturbations (CSPs) to map the binding pocket of the methyllysine-binding domains and determine dissociation constants. We do not describe the procedures for NMR data collection and processing or for obtaining resonance assignments here, which should be pursued with the help of an expert.

• (a)Samples should be prepared in buffer suitable for NMR spectroscopy. It is generally advisable to have a buffer with a pH ≤7.0, total salt concentration of ≤150 mM and free of any viscous additives, such as glycerol, to obtain the best signal. Deviation from these optimal conditions is possible with careful experimental planning and should be discussed with an expert. Reducing agents, such as DTT, are fine to use here (of note, for some NMR experiments deuterated DTT is required). For the following analysis, the methyllysine recognition domain should be enriched in ¹⁵N and the final sample should be ~250–300 μL in volume with a concentration of ~100 μM (assuming the use of a shigemi NMR tube). Lower concentrations can be used if a cryogenic probe is available. Peptide stock should be prepared in matching buffer. This is critical as NMR resonances are sensitive to buffer conditions.
• (b)Collect an ¹H,¹⁵N-heteronuclear single quantum coherence (¹H,¹⁵N HSQC) spectrum of the apo-state of the protein on a spectrometer of ≥500 MHz. Confirm that conditions chosen provide a good signal to noise ratio and that the majority of expected resonances are observed.
• (c)Collect additional ¹H,¹⁵N HSQC spectra after adding increasing concentrations of the unlabeled methyllysine substrate to the NMR sample until saturation is reached. For very weak-binding substrates, it may not be feasible to reach saturation. Overlay of these spectra will reveal resonance perturbations induced by ligand binding (Fig. 5).
• (d)
  CSPs can be determined by computing the change in ¹H and ¹⁵N chemical shifts, Δδ_H and Δδ_N, where Δδ = δ_bound - δ_apo. These values can be normalized using the equation:

  $$\mathrm{\Delta }{\delta }_{\text{norm}}={\left(\mathrm{\Delta }{\delta }_{\text{H}}{}^2+{\left(\mathrm{\Delta }{\delta }_{\text{N}}/5\right)}^2\right)}^{1/2}$$

Fig. 5.

Overlay of successive ¹H,¹⁵N HSQC spectra of a labeled methyllysine reader domain upon titration of methylated peptide. Arbitrary peptide concentration is shown in legend.

• (e)If resonance assignments have been made, they can be used to determine the residues directly or indirectly involved in binding of the methylated substrate, which can be further confirmed via mutagenesis.
• (f)Binding constants can be obtained by nonlinear fitting of Δδnorm vs ligand concentration using an equation that accounts for ligand depletion. Note that given the high protein concentration utilized in the NMR experiment, it is advisable to determine binding constants of K_d < 10 μM using an alternative approach as outlined above, as these K_ds will lead to stoichiometric binding conditions in the NMR experiment.

6. STRUCTURE DETERMINATION BY X-RAYCRYSTALLOGRAPHY

X-ray crystallography is the most widely utilized technique for determining the structure of methyllysine reader domains in complex with their cognate substrates. Obtaining well diffracting crystals is an empirically driven process that can take considerable time and effort. Through ample screening one can obtain the appropriate chemical, biochemical, and physical conditions that will produce a crystalline state of the complex as a starting point. From here extensive optimization of conditions might yield diffracting crystals, from which a suitable data set can be obtained and a structure determined. Below is a set of guidelines to aid in obtaining crystals. The process of data collection, processing, and structure determination is not covered here and should be pursued with the help of an expert. Note that the phase problem for most methyllysine reader domains can be determined through molecular replacement using previously solved homologous domains as a model, or in the case of Zn-binding domains can be determined de novo using MAD or SAD.

• (a)Beginning with the most minimal construct is critical in the crystallization process and should be considered for both the reader domain and the substrate peptide. If the construct contains significant regions outside those necessary for binding, especially if these regions are not structured, this can preclude crystallization.
• (b)A highly pure and homogenous sample is vital in the crystallization process, and thus extensive purification of the protein should be pursued before screening begins. In addition, having freshly purified and not previously frozen samples often yields better results.
• (c)The protein should be as concentrated as possible while retaining stability as a starting point.
• (d)Cocrystallization is generally the best strategy for methyllysine reader domains. Incubate the protein with a slight molar excess of peptide (1.2–1.5:1.0 molar ratio of peptide:protein) for at least 1 h prior to setting up crystal trays. If the complex has a high enough affinity, it can be repurified using size-exclusion chromatography to obtain a highly homogenous stock mixture. However, weaker affinity complexes will likely partly dissociate during this process, and thus the protein–peptide mixture should be used directly after incubation.
• (e)A large number of “shotgun” screens are commercially available and provide the best starting conditions if nothing is known about the crystallization properties of the domain of interest. These will allow for testing a wide range buffers, salts, and precipitants. It is also advisable to screen for different crystallization temperatures.
• (f)Once a promising condition is found, systematic screening can be pursued to obtain the best diffracting crystals.