Peptide arrays synthesized by the SPOT method can be used to analyze the substrate specificity of Protein lysine methyltransferases (PKMTs) and to define the substrate spectrum of PKMTs to understand their biological role. This protocol describes how to synthesize peptide arrays, methylate them with PKMTs, and analyze the results.
Lysine methylation is an emerging post-translation modification and it has been identified on several histone and non-histone proteins, where it plays crucial roles in cell development and many diseases. Approximately 5,000 lysine methylation sites were identified on different proteins, which are set by few dozens of protein lysine methyltransferases. This suggests that each PKMT methylates multiple proteins, however till now only one or two substrates have been identified for several of these enzymes. To approach this problem, we have introduced peptide array based substrate specificity analyses of PKMTs. Peptide arrays are powerful tools to characterize the specificity of PKMTs because methylation of several substrates with different sequences can be tested on one array. We synthesized peptide arrays on cellulose membrane using an Intavis SPOT synthesizer and analyzed the specificity of various PKMTs. Based on the results, for several of these enzymes, novel substrates could be identified. For example, for NSD1 by employing peptide arrays, we showed that it methylates K44 of H4 instead of the reported H4K20 and in addition H1.5K168 is the highly preferred substrate over the previously known H3K36. Hence, peptide arrays are powerful tools to biochemically characterize the PKMTs.
In the last two decades, several reports demonstrated the importance of post-translational modifications (PTM) in cellular development and several diseases like cancer, but recently protein lysine methylation has emerged as an another vital PTM. While initially histone lysine methylation was found to be an essential chromatin mark, later work also showed lysine methylation of several non-histone proteins 1-4. The sequential transfer of methyl groups from S-adenosyl-L-methionine to the ε-amino group of lysine residues is catalyzed by a family of enzymes called Protein Lysine Methyltransferases (PKMTs) that contains over 60 proteins in the human genome. PKMTs were initially discovered as histone modifying enzymes that methylate specific lysine residue but later reports demonstrated that they could also methylate non-histone proteins 5. Up to now, approximately 5,000 lysine methylation sites were identified on different proteins 6, but the enzymes responsible for these modifications are often not identified. One reason for this is that the specificity of most PKMTs has not been studied extensively. Therefore, it recurrently occurs that novel substrates of PKMTs are discovered. The lack of a detailed knowledge of the substrate specificity of PKMTs hinders understanding of their biological function and role. To study the specificity of a PKMT in detail, the methylation rates of many peptide substrates that differ in one or few amino acids must be measured and compared, which is ideally done using peptide arrays. Based on the resulting specificity profiles, potential substrates of PKMTs can be identified that can be studied further.
Peptide arrays are widely used tools for the biochemical analysis of antibodies, peptide modifying enzymes and mapping of protein-protein interaction sites (antibody-antigen, receptor-ligand) 7-9. Several hundreds of peptides are needed for such applications. Different methods are available for peptide synthesis, among them peptide synthesis on resin is very commonly used, but it has limitations in throughput and it is relatively expensive. These issues were resolved with the introduction of the SPOT synthesis method by Frank and colleagues 10. The SPOT synthesis method allows synthesis of several hundred peptides in parallel and on average it is inexpensive compared to resin synthesis. The peptides synthesized on cellulose membrane can be used either directly for various applications or peptides can be cleaved from the membrane and used as free peptides for in solution assays or to prepare peptide microarrays 10-13.
SPOT synthesis is a variant of the solid phase peptide synthesis, which uses a cellulose membrane as a solid support and employs the standard Fmoc-chemistry 10-13. Hence, the synthesis of peptide chains starts at the C-terminal end and proceeds towards the N-terminal end in contrast to the biological synthesis in ribosomes. Cellulose membranes are functionalized for the attachment of the first activated amino acids (Fig. 1). The SPOT method is based on the sequential delivery of activated amino acids in a droplet of solvent to defined spots on the membrane using an automated pipetting system. The droplet of liquid is dispensed on the porous membrane where it forms a circular wet spot, which later acts as an open reactor for the chemical reactions in peptide synthesis. The spot size is determined by the volume dispensed and the absorptive capacity of the membrane, multiples of such spots are arranged as arrays. The scale of synthesis correlates with the spot size and the loading capacity of the membrane. The distance between the spots and the density of arrays are managed by varying the spot sizes. Cellulose membrane has several advantages as solid phase in peptide synthesis, it is inexpensive, tolerant to the chemicals used in peptide synthesis, stable in aqueous solutions and easy to handle. In addition, its hydrophilic nature makes it suitable for several biological assay systems. SPOT synthesis can be carried out manually or automated (for 1000s of peptides) depending on the required number of peptides. A fully automated SPOT synthesizer from Intavis (Köln, Germany) is used for our applications. It permits synthesis of peptides in different amounts and of different length. Linear peptides are regularly synthesized with 15 to 20 amino acid length, in addition peptides of up to 42 amino acid can also be prepared by step-wise synthesis 14,15. However, increasing the number of amino acids leads to reductions in the overall coupling yields, which affects the quality of the peptides. Because of the low amount of peptides per spot, the products are often difficult to purify and the quality of individual peptides cannot be easily assessed. Therefore, the results obtained from SPOT peptide arrays must be confirmed either with peptides synthesized by standard methods in larger scale, which can be purified and analyzed according to standards in peptide synthesis or by synthesizing the proteins containing the desired peptide sequences. Still, we found the SPOT synthesis to be highly reliable and results generally to be reproducible. SPOT synthesis is not restricted to proteinogenic amino acids, several commercially available modified amino acids also can be used for synthesis, allowing peptides to be modified before and after the final cleavage of the side-chain protection group and, furthermore, it also allows incorporation of phosphorylated, methylated or acetylated amino acids 11.
Immobilized peptide libraries synthesized by the SPOT method can be directly used for many biological and biochemical assays. We employed peptide arrays comprising 300-400 peptides to investigate the substrate specificity of PKMTs. For enzymatic modification, the peptide arrays are incubated with the respective PKMT and labeled [methyl-3H]-AdoMet in an appropriate buffer. The methylation of the respective substrate is analyzed by following the enzymatic transfer of the radioactively labelled methyl groups from AdoMet to the peptide substrate via autoradiography (Fig. 3). By this procedure, the peptide arrays allow the study of methylation of different peptide substrates at the same time. One important advantage of this method is that all the peptides are methylated in competition, such that during the linear phase of the methylation kinetics, the relative methylation of each peptide is proportional to the catalytic rate constant divided by the dissociation constant (kcat/KD) of the enzyme for the respective peptide substrate. Therefore, the amount of radioactivity incorporated into each spot is directly correlated with the enzymatic activity towards the particular peptide. Using the results of a peptide array methylation experiment, the specificity profile of the PKMT can be defined and based on this novel substrates can be predicated. Peptide arrays allow the rapid and cost efficient validation of the methylation of novel substrates at the peptide level. For this, arrays are prepared that contain the predicted novel substrates together with modified peptides containing an Ala instead of Lys at the target sites as well as positive and negative control peptides. Finally, the novel substrates can be prepared as proteins together with mutants, in which the target Lys is altered to Ala and the methylation can be confirmed at the protein level. Depending on the results, this is then followed by biological studies addressing potential roles of the methylation of the newly described protein substrates.
1. Preparation of Peptide Arrays
2. Protein Expression and Purification
3. Peptide Array Methylation
4. Data Processing and Analysis
Peptide arrays were successfully used to biochemically characterize the specificity of PKMTs, and several novel histone and non-histone substrates of PKMT were identified by this approach 5,16-19. Defining the correct substrate spectrum of a PKMT (or any enzyme) is an essential step towards understanding of its molecular mechanism and cellular functions.
As an example of the application of the peptide SPOT array methylation method, we describe the results of the specificity analysis of the NSD1 PKMT 19. Previous to our work, this enzyme was described to methylate several substrates including histone H3 at K36 and histone H4 at K20 but also the NF-kB family transcription factor p65 at K218 and K221. Fig. 4A shows an example of a methylation reaction of a peptide array with NSD1. For this experiments, a large peptide array based on the H3 (31-49) sequence was synthesized that contains all possible single amino acid alterations of the original sequence to test the significance of each amino acid for NSD1 interaction and methylation. In total 380 peptides were synthesized (20 possible amino acids x 18 residues plus one original H3 sequence in each row). The horizontal axis represents the sequence of the peptide and in the vertical direction the amino acid which is altered in the corresponding peptide is indicated. For example, the spot at line 17 column 4 contains a threonine at third position instead of glycine that is present in the wild type sequence (Fig. 4A). In such a way, point mutations are generated to test the preference of NSD1 for each native amino acid at each position in the peptide substrate. Methylation with NSD1 showed that it specifically acts on H3K36 and it has preferences on either side of the target lysine from 34 to 38.
Figure 4B represents the consolidation of the three peptide array methylation experiments with NSD1. The quantitative information was acquired from the individual peptide arrays and the results were normalized and averaged as described above. The standard deviations of individual averages show that the data are highly reproducible. Overall around 85% of the peptides shows SDs smaller than 20% and more than 97% of the peptide substrates demonstrated SDs smaller than 30% (Fig. 4C). In addition, we also calculated the discrimination factor to precisely determine the contribution of each amino acid at the tested positions. As described it provides the quantitative description of the peptide read out and preference of amino acids at the specific position (Fig. 5A). Our data show that NSD1 prefers aromatic residues at the -2 position (considering K36 as position 0) (F > Y > G); hydrophobic residues at -1 (I > L > V); basic residues at +1 (R > QKNM), where it cannot tolerate hydrophobic or aromatic residues; and hydrophobic residues at +2 (V > IA > P). At other sites (like -3, +3, or +6), NSD1 prefers some amino acids, but no strong residue specific readout was detected.
Based on this profile, potential novel peptide substrates can be found by database searches like using Scansite 20 (Fig. 5B). These peptides were prepared on a SPOT array including positive and negative controls and incubated with NSD1 and radioactively labelled AdoMet to identify the subset of them that are methylated at peptide level (Fig. 5C). As follow ups, peptide arrays can be prepared which include peptide variants, in which the target lysine is replaced by alanine, to confirm that methylation takes place at the predicted site. Then, the target proteins or subdomains of them containing the target lysine can be produced recombinantly and the methylation also tested at protein level. Finally, depending on the results, follow up experiments can investigate if the methylation also occurs in cells and which biological role it has.
Figure 1: Scheme of peptide synthesis by the SPOT method on cellulose membrane. Please click here to view a larger version of this figure.
Figure 2: Example of a peptide array stained with Bromophenol blue to confirm the synthesis of peptides. Please click here to view a larger version of this figure.
Figure 3: Scheme of the peptide array methylation experiment; peptide arrays were methylated by incubation with PKMT and labeled [methyl-3H]-AdoMet in an appropriate buffer. Afterwards the radioactivity transferred to each spot is detected by autoradiography. Please click here to view a larger version of this figure.
Figure 4: Example results obtained with the NSD1 PKMT. A) Example of a substrate specificity peptide array for NSD1 using the H3 (31-49) sequence as template. The horizontal axis represents the H3 sequence and the vertical axis represents the amino acids by which the corresponding row is mutated. The first row contains peptides with original sequence. B) Consolidation of results from 3 independent peptide array methylation experiments with NSD1, the data were averaged results from all three experiments after normalization. C) Distribution of standard errors for the average methylation data shown in panel B. Reproduced from Kudithipudi et al. (2014) with some modifications 19. Please click here to view a larger version of this figure.
Figure 5: Discovery of novel PKMT substrates. A) Discrimination factors of NSD1 for the recognition of amino acid residues at the positions next to the target lysine (K36 in the experiment shown here). The data show that NSD1 prefers aromatic residues at the -2 position (considering K36 as position 0). Hydrophobic residues are recognized at the -1 and +2 positions, albeit with distinctive differences in details. At the +1 site basic residues and amides are preferred. At other sites some weak preferences, but no strong residue specific readout was detected. B) Screenshot of an example search at Scansite, using the specificity profile determined for NSD1. C) Peptide SPOT array containing several predicted novel NSD1 substrates together with positive (H3K36) and negative controls (H3K36A). Some of the predicated novel targets were strongly methylated (some of them annotated), in other cases the predictions could not be verified. Panel A and C is reproduced from Kudithipudi et al. (2014) with some modifications 19. Please click here to view a larger version of this figure.
Base | 1 M Oxyma Pure in NMP -> 2,13 g Oxyma Pure in 15 ml NMP |
Activator | 2.4 ml N,N’-Diisopropylcarbodiimide in 15 ml NMP |
Capping Mixture | 20% acetic anhydride in DMF -> 6 ml Acetic anhydride in 30 ml DMF |
Fmoc Deprotection | 20% Piperidine in DMF -> 200 ml in 1,000 ml DMF |
Sidechain Deprotection | 900 µl Triisopropylsilane + 600 µl ddH2O in 30 ml TFA |
Staining | 0.02% bromophenol blue in DMF |
Table 1: Chemical mixtures and their composition used in the protocol.
SPOT synthesis as described here is a powerful method to map protein-protein interaction sites and investigate the substrate recognition of peptide modifying enzymes. However, SPOT synthesis still has certain drawbacks, because although the peptides synthesized by the SPOT method are reported to have more than 90% purity 21 this is difficult to confirm in each instance. Therefore, results have to be reproduced by other methods for instance using protein domains or with purified peptides synthesized by standard methods. In addition, the other major drawback of SPOT arrays is that most time they cannot be reused to carry out several assays with one array.
To increase the efficiency of synthesis with amino acids of low stability, like the protected arginine, it is important to make them fresh for every 5 cycles. In enzymatic reactions or protein binding studies it is often observed that the periphery of the peptide spot has more signal intensity than the center (“ring spot effect”). This effect can be due to the high density of peptides in the central part of the spot and then it can be resolved by downscaling the peptide synthesis 22. For further trouble shooting of peptide synthesis the machine handbook and company service should be consulted. If methylation is not observed, positive control peptides (i.e. peptides known to be methylated by the enzyme under investigation) should be added to the membrane.
Alternate to SPOT arrays, high density peptide micro arrays 23 are available, but they are more expensive and require special equipment for usage. Furthermore the amount of peptide per spot is smaller, which requires more sensitive readout procedures. Protein arrays are also available to study these functions 24,25, but they are more difficult to prepare because each individual protein needs to be expressed and purified and protein folding on the array needs to be maintained. An additional advantage of peptide arrays is the possible introduction of post- translational modifications during the synthesis and the ease of mutational testing of the roles of individual amino acid residues.
It is an essential starting point for this protocol, to have at least one peptide identified, which is methylated by the enzyme under investigation. Methylation conditions and buffers have to be optimized, as well as the concentration of the methyltransferase. Exemplary time courses of methylation have to be determined to avoid complete methylation of the best substrate, which will cause loss of dynamic range.
The authors have nothing to disclose.
This work has been supported by the DFG grant JE 252/7.
Name of Material/ Equipment | Company | Catalog Number | Comments/Description |
Ethanol abs Gradient Grade HPLC | Honeywell | 10299901 | Flammable |
N,N-Dimethylformamide Peptide Synthesis | Biosolve | 4193302 | Flammable, Toxic |
Piperidine ≥ 99 % for Peptide Synthesis | Roth | A122.1 | Flammable, Toxic, corrosive |
Oxyma Pure (Ethyl (hydroxyimino)cyanoacetate ) | Novabiochem | 8510860100 | |
N,N’-Diisopropylcarbodiimide purum ≥ 98% GC | Fluka | 38370 | Flammable, Toxic, corrosive |
Acetic anhydride | Roth | CP28.1 | Flammable, Toxic, corrosive |
Triisopropylsilane 99%, | Aldrich | 233781 | Flammable, Toxic |
Dichlormethane ≥99,9% | Roth | P089.1 | Carcinogenic |
N-Methyl-2-Pyrrolidone | Roth | 4306.2 | Toxic |
Derivatized cellulose Membrane | Intavis AG Köln | 32.1 | |
Trifluoroacetic acid | Roth | P088.2 | Toxic, corrosive |
Bromphenolblue | AppliChem | A3640.0010 | |
Ammonium Hydrogen Carbonate | Roth | T871.2 | Toxic |
Sodium Dodecyl Sulfate Pellets | Roth | CN30.3 | Toxic, Flammable |
MultiPep Synthesizer | Intavis AG Köln | n.a. | |
HyperfilmTM high performance film | GE Healthcare | 28906837 | |
Phoretix software | TotalLab | n.a. |