Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Am Chem Soc. Author manuscript; available in PMC Jan 23, 2014.
Published in final edited form as:
PMCID: PMC3582175
Profiling Genome-wide Chromatin Methylation with Engineered Posttranslation Apparatus within Living Cells
Rui Wang,1,2§ Kabirul Islam,1§ Ying Liu,3 Weihong Zheng,1 Haiping Tang,4 Nathalie Lailler,5 Gil Blum,1,6 Haiteng Deng,4 and Minkui Luo1*
1Molecular Pharmacology and Chemistry Program, Memorial Sloan-Kettering Cancer Center, New York, NY 10065
2Program of Pharmacology, Weill Graduate School of Medical Science, Cornell University, New York, NY 10021
3Program of Biochemistry & Structural Biology, Cell & Developmental Biology, and Molecular Biology, Weill Graduate School of Medical Science, Cornell University, New York, NY 10021
4School of Life Sciences, Tsinghua University, Beijing, 100084 China
5Genomics Core Laboratory, Memorial Sloan-Kettering Cancer Center, New York, NY 10065
6Tri-Institutional Training Program in Chemical Biology, Memorial Sloan-Kettering Cancer Center, New York, NY 10065
Corresponding Author Minkui Luo: luom/at/
§These authors made equal contribution.
Protein methyltransferases (PMTs) have emerged as important epigenetic regulators in myriad biological processes both in normal physiology and disease conditions. However, elucidating PMT-regulated epigenetic processes has been hampered by ambiguous knowledge about in vivo activities of individual PMTs particularly because of their overlapping but non-redundant functions. To address limitations of conventional approaches in mapping chromatin modification of specific PMTs, we have engineered the chromatin-modifying apparatus and formulated a novel technology, termed Clickable Chromatin Enrichment with parallel DNA sequencing (CliEn-seq), to probe genome-wide chromatin modification within living cells. The three-step approach of CliEn-seq involves in vivo synthesis of S-adenosyl-L-methionine (SAM) analogues from cell-permeable methionine analogues by engineered SAM synthetase (methionine adenosyltransferase or MAT), in situ chromatin modification by engineered PMTs, subsequent enrichment and sequencing of the uniquely modified chromatins. Given critical roles of the chromatin-modifying enzymes in epigenetics and structural similarity among many PMTs, we envision that the CliEn-seq technology is generally applicable in deciphering chromatin methylation events of individual PMTs in diverse biological settings.
Enzymatic transfer of the methyl group from the cofactor S-adenosyl-L-methionine (SAM) to DNA, RNA, protein and metabolite substrates accounts for the repertoire of biologically relevant methylation.1 As an important subtype, protein posttranslational methylation increases functional diversity of proteome and thus influences numerous aspects of normal cell biology.2 In contrast, misregulation of these events has been implicated in many diseases including developmental disorders and cancer.2 The human genome encodes more than 60 protein methyltransferases (PMTs) and their epigenetic functions are mainly executed via methylating diverse histone and nonhistone targets.3
PMT-mediated histone methylation, together with other chromatin events such as DNA methylation and ATP-dependent chromatin remodeling, establishes specific chromatin states for gene regulation, chromosome recombination and DNA repair.4,5 To elucidate biological roles of individual PMTs on these chromatin events, it is critical to define their genome-wide chromatin modification sites in an unambiguous manner. However, considerable challenge exists for conventional approaches to define such sites for a single PMT because PMTs often harbor overlapping chromatin-modifying activities (e.g. the methylation of histone H3 lysine 9 by G9a, GLP1, SETDB1/2, SUV39H1/2 and PRDM2).6-9 Moreover, dynamic chromatin modifications often need to be executed via multiple cellular machineries within living cells.4 To address these challenges in deciphering the activities for individual PMTs in chromatin landscape, we envisioned an approach suitable for the in-cell setting.
In living cells, a native methylation reactome consists of two crucial enzymatic reactions (Figure 1): SAM biosynthesis and substrate labeling. The biosynthesis of SAM 2a from methionine 1a and adenosine triphosphate (ATP) is carried out by methionine adenosyltransferases (MATs).10 Here we envisioned in-cell-based mapping of chromatin methylation using an engineered posttranslation reactome. This approach is featured by using an engineered methionine adenosyltransferase (MAT) within living cells to process membrane-permeable methionine analogues into SAM analogues, which themselves show limited cell-permeability.11 The in-cell-produced SAM analogues can then be utilized in situ by compatible PMT variants to label chromatin with distinct chemical handles for target enrichment and characterization (Figure 1).
Figure 1
Figure 1
Schematic description of profiling genome-wide chromatin methylation with a native or engineered posttranslation apparatus within a living cell. In a native context, methionine 1a and adenosine triphosphate (ATP) will be processed by methionine adenosyltransferases (more ...)
G9a (EuHMT2/KMT1C) and GLP1 (Eu-HMT1/KMT1D) are two human oncogenic protein lysine methyltransferases (PKMTs) with their epigenetic functions implicated in silencing viral genes and tumor suppressors.12 G9a and GLP1, together with other H3K9 PMTs (e.g. SUV39H1, SETDB1 and PRDM2), methylate H3 lysine 9 (H3K9) via multimeric complexes in cellular contexts.6-9 Dissecting chromatin-modifying activities of these H3K9 PMTs is challenging but important toward elucidating their physiological and pathogenic functions. With G9a and GLP1 as representative PMTs, we demonstrated the use of methionine analogues, a MAT variant and engineered PMTs for substrate labeling in vitro and genome-wide chromatin modifications within living cells.
Engineering MATs for enzymatic synthesis of SAM analogues in vitro
In the course of developing in vitro approaches to profile protein methylation, terminal alkyne-containing (E)-pent-2-en-4-ynyl-SAM (EnYn-SAM, 2b) and (E)-hex-2-en-5-ynyl-SAM (Hey-SAM, 2c) have been characterized as active cofactors of G9a’s Y1154A mutant and used to label G9a’s targets in vitro.13 To transform the approach to living cells, we first examined the feasibility of MAT-mediated chemoenzymatic synthesis of the two SAM analogues from the corresponding methionine analogues.
In a biological setting, SAM is synthesized from endogenous ATP and cell permeable methoinine by highly conserved MATs.14,15 Upon analyzing the structures of MATs (PDB code 1RG9 for E. Coli MAT10 and PDB code 2P02 for human MAT2A), we noticed the existence of a highly-conserved small methylthio-binding pocket formed by hydrophobic I117, G120, V121 and I322 residues (the sequence number of human MAT2A, Figure 2a). To expand the pocket to accommodate bulky methionine analogues such as S-(E)-pent-2-en-4-ynyl and S-(E)-hex-2-en-5-ynyl homocysteine (EnYn-Met 1b and Hey-Met 1c, Figures 1, ,2b,2b, Experimental Method and Supporting Information), the I117, V121 and I322 residues were systemically replaced with smaller amino acids (Ala, Gly, Leu or Val) individually or in a combination manner. The active MAT2A variants were identified upon coupling with G9a’s Y1154A mutant and then detecting the anticipated alkylation product in the presence of ATP and EnYn-Met 1b or Hey-Met 1c. Among the 17 examined MAT2A mutants, the I117A variant has the best activity towards EnYn-Met 1b and Hey-Met 1c as shown by its highest alkylation efficiency (30% for 1b and 100% for 1c, Figures 2b-d, Figure S1). In contrast, native MAT2A and other mutants are either inert or slightly active towards the methionine analogues (Figures 2b, S1). The general preference of Hey-Met 1c over EnYn-Met 1b revealed in the two-enzyme coupled assays might attribute to the higher activities of the MAT2A variant on Hey-Met 1c or the G9a mutant on Hey-SAM 2c, which will be examined elsewhere. Here the efficient conversion of Hey-Met 1c into Hey-SAM 2c by the I117A mutant presents the MAT2A variant as a suitable component to construct a bioorthogonal posttranslation apparatus.
Figure 2
Figure 2
Engineering MAT2A for chemoenzymatic synthesis of SAM analogues. (a) Key residues consisting of the hydrophobic methylthio-binding pocket in E. Coli MAT (PDB 1RG9) and human MAT2A (PDB 2P02). The backbone structure was constructed on the basis of E. Coli (more ...)
Generality of engineered MATs coupled with PMT variants for in vitro substrate labeling
To examine the generality of the MAT2A I117A mutant, we explored whether the MAT variant is compatible with other PMT variants. G9a-like protein 1 (GLP1) is a human protein lysine methyltransferase, which also acts on H3K9.12 The SAM-binding motif of GLP1 can be tailored (GLP1-Y1211A) to incorporate the bulky SAM analogue Hey-SAM 2c (manuscript in revision). Besides PKMTs, protein arginine methyltransferase 1 (PRMT1), the predominant mammanlian type 1 PRMT,16 was also successfully engineered and its Y39FM48G variant has been shown to be active towards another SAM analogue 4-propargyloxy-but-2-enyl SAM (Pob-SAM).17 Upon coupling the GLP1 variant with the MAT2A I117A mutant in the presence of Hey-Met 1c and ATP, a robust alkylation of GLP1’s H3K9 peptide substrate was observed (Figure S1). In a similar manner, the MAT2A I117A mutant can process S-4-propargyloxy-but-2-enyl homocysteine (Pob-Met) into Pob-SAM and, upon coupling with the PRMT1 Y39FM48G mutant, to label the PRMT1 substrate RGG peptide (here Pob-SAM but not Hey-SAM 2c was found to be an active cofactor of the PRMT1 mutant in vitro17) (Figure S2). The coupled activities of the MAT mutant and the engineered G9a/GLP1 variants were also validated with the full-length H3, an in vivo protein substrate of G9a and GLP1.6 Tandem MS of the tryptic H3 peptides confirmed the expected (E)-hex-2-en-5-ynylation on H3K9, the known methylation site of the two PMTs (Figure S3). The MAT2A I117A variant thus demonstrated the robust compatibility with multiple bulky methionine analogues and engineered PMTs for in situ generation of SAM analogues.
Biosynthesis of SAM analogues within living cells
SAM analogues show poor membrane permeability and thus is restricted for the access to intracellular PMTs.11 In addition, chemically-synthesized SAM analogues consist of diastereomeric mixtures with the only sulfonium-(S)-epimer bioactive.14 To circumvent these limitations, we envision that, in the presence of the engineered MAT in living cells, pure bioactive SAM analogues can be produced in situ from the corresponding membrane-permeable methionine analogues. With the MAT2A I117A variant characterized above, we investigated its activity to process Hey-Met 1c into Hey-SAM 2c in living cells. Here the MAT2A I117A mutant was expressed in human embryonic kidney 293T (HEK293T) cells followed by the supply of Hey-Met 1c in a methionine-free growth medium. Hey-SAM 2c was readily extracted from the cell pellets (Figure 3a) and confirmed by MS (Figure S5), consistent with the intracellular production and maintenance of Hey-SAM 2c. Hey-SAM’s in-cell biosynthesis was further confirmed by the dose-dependent increase of intracellular Hey-SAM 2c, which reaches a plateau around 1 mM Hey-Met 1c (ESI-MS, Figures 3a-b, S5). In contrast, no Hey-SAM 2c was detected when cells were transfected with native MAT2A or grown in the absence of Hey-Met 1c (Figures 3a, S4a), emphasizing the roles of the MAT2A I117A mutant and Hey-Met 1c for intracellular production of Hey-SAM 2c.
Figure 3
Figure 3
Biosynthesis of Hey-SAM 2c and Hey-SAM-mediated histone alkylation within living cells. (a) Production of Hey-SAM 2c within living cells. HEK293T cells were transfected with the MAT2A I117A and supplied with various concentrations of Hey-Met 1c. The resultant (more ...)
To examine the efficiency of engineered MAT2A to produce Hey-Met 1c within living cells, we quantified the amount of intracellular Hey-SAM 2c by LC/MS (see Experimental Section for details). The presence of 1 mM Hey-Met 1c in the growth medium led to the intracellular accumulation of approximately 4.7 nmol Hey-SAM/106 cells, which is comparable with the reported amount of SAM in cells (~5 nmol SAM/106 cells).18 Moreover, the toxicity of Hey-Met 1c was evaluated with the alamar blue assay. The presence of < 2 mM Hey-Met 1c has no obvious effect on the viability of HEK293T cells. When the cells were transfected with the MAT2A I117A plasmid, only slight growth inhibition (~15% compared to control) was observed in the presence of Hey-Met 1c (Figure S7). These findings suggest that neither Hey-Met 1c nor Hey-SAM 2c has significant cellular toxicity under the examined concentrations. These results collectively argue that Hey-Met 1c can penetrate through cell membranes and, together with endogenous ATP, be processed into Hey-SAM 2c by the MAT2A I117A variant within living cells. The intracellular production of the SAM analogue thus circumvents its membrane permeability issue,11,14 and enables its utility within living cells.
Using engineered methylation apparatus to label targets of G9a and GLP1 within living cells
Given that both G9a’s Y1154A mutant and GLP1’s Y1211A mutant can process Hey-SAM 2c in vitro for substrate labeling, we then constructed the engineered posttranslation apparatus within living cells by co-expressing the MAT2A I117A variant with the two PMT mutants, respectively. After growing the cells in the presence of Hey-SAM’s biosynthetic precursor Hey-Met 1c (1 mM), cell lysates were subject to the click chemistry with a cleavable azido-diazo-biotin probe (see Experimental Method and Supporting Information and the protocol below for profiling genome-wide chromatin modification).19 The putative (E)-hex-2-en-5-ynylated proteins were enriched with streptavidin beads, cleaved from streptavidin beads after washing and then blotted with anti-H3 antibody (Experimental Method and Supporting Information). Histone H3 was pulled down only from the samples containing the engineered MAT, the PMT variants and Hey-Met, consistent with the in vivo activity of native G9a and GLP1 on histone H3. In contrast, no labeling was observed for the controls that lack any of the three components or were transfected with native MAT, G9a or GLP (Figures 3b, S4c), indicating the prerequisite of the engineered components (MAT/G9a/GLP1 variants and the matched Met analogues) for the bioorthogonal target labeling within living cells. To further verify that the PMT variants, like native G9a and GLP1, act on H3K9 within living cells, the total histones were acid-precipitated and analyzed by LC-MS/MS. The desired (E)-hex-2-en-5-ynylation on lysine 9 in histone H3 demonstrated that the engineered apparatus recapitulates the activities of native G9a and GLP1 for target labeling within living cells (Figure S6).
To examine the efficiency of the engineered methylation apparatus, we implemented this method to probe the time-dependent global methylation activity of G9a within living cells. The cells were transfected with the MAT2A I117A and G9a Y1154A mutants and then treated with Hey-Met 1c (1mM) for a varied period of time. The cell lysates were recovered and reacted with an azido fluorescent dye via copper-catalyzed azide–alkyne cycloaddition (CuAAC) reaction, followed by visualization with in-gel fluorescence. As revealed previously that G9a acts on histone and diverse nonhistone targets, the labeling of G9a targets can be readily observed after 8 h treatment of Hey-Met 1c to the cells carrying the engineered MAT2A and G9a mutants. In contrast, only marginal background labeling was observed in the cells transfected with the MAT2A I117A mutant alone (Figure S9), arguing the importance of the presence of the engineered G9a for the targeting labeling. The background labeling might be due to residual activities of certain native PMTs on Hey-SAM 2c. Given that histone demethylases have high specificity for substrates,20,21 the bulky (E)-hex-2-en-5-ynylated proteins may not be substrates of wild type demethylases and thus resistant to dealkylation. The issues of the background and stability of the (E)-hex-2-en-5-ynylated proteins remain to be addressed with additional experiments.
Profiling genome-wide chromatin-modifying activities of G9a and GLP1
With the feasibility to construct the in-cell-based bioorthogonal posttranslation apparatus, we further leveraged the system to examine genome-wide chromatin-modifying activities of G9a and GLP1 via an unprecedented approach, termed as Clickable Chromatin Enrichment with parallel DNA sequencing (CliEn-seq). Here the (E)-hex-2-en-5-ynylated chromatins, generated by the engineered MAT2A and G9a/GLP within the living cells supplied with Hey-Met 1c, were further conjugated to the cleavable azido-diazo-biotin probe via CuAAC,19 followed by affinity pull-down with streptavidin beads (Figure 4a). We observed the 30~50-fold enrichment of chromatin DNA from the cells co-transfected with engineered MAT2A I117A and G9a Y1154A/GLP1 Y1211A in contrast with the control group transfected with the MAT2A I117A mutant alone (Figure 4b). To resolve the gene loci containing the (E)-hex-2-en-5-ynylated chromatins, the enriched chromatin DNA was subject to next-generation sequencing. The CliEn-seq data suggest that the activity of G9a is broadly spread around genome-wide chromatins in HEK293T cells including the previously-known G9a-targeted loci myoD, rb1 and bmi1 (Figures 4c, S10).22,23 The enriched activity of G9a at these representative loci was independently confirmed by qPCR (CliEn-qPCR, Figures 4d, S10).
Figure 4
Figure 4
Profiling genome-wide chromatin modification of G9a and GLP1. (a) Schematic description of Clickable Chromatin Enrichment with parallel DNA sequencing (CliEn-seq). Here the terminal-alkyne-modified chromatins were conjugated with the cleavable azido-diazo-biotin (more ...)
G9a and GLP1 were proposed to account for the activities of methylating histone H3 (H3K9me2) and nonhistone targets, respectively.24 Remarkably, our CliEn-seq and CliEn-qPCR data showed that G9a and GLP1 harbor complementary chromatin-modifying activities at the loci, which were previously known to be G9a targets (Figures 4c-d, S10). This finding is also consistent with our chromatin immunoprecipitation (ChIP) experiment using anti-H3K9me2 antibody, where overexpressing either G9a or GLP1 enhanced the H3K9me2 marker on the targeted promoters of the examined genes (Figure S10). To further distinguish the activities of G9a/GLP1 between the promoter regions and gene bodies, we selected 35 gene loci known to contain H3K9 methylation sites (Table S2) and dissected the CliEn-seq data of the (E)-hex-2-en-5-ynylation pattern according to their translation start sites (TSS) and translation end sites (TES). The comparable enrichment of the G9a- or GLP1-mediated (E)-hex-2-en-5-ynylation across TSS and TES suggests that G9a and GLP1 have no significant preference between the gene promoters and the gene bodies in HEK293T cells (Figure 5). The CliEn-seq approach using the engineered posttranslation system thus enables to decipher the activities of closely-related PMTs within living cells.
Figure 5
Figure 5
CliEn-seq signal enrichment profile near the transcription start site (TSS) (a) and the transcription end site (TES) (b) for 35 selected genes with known H3K9 methylation markers (see Supporting Information for the list of the genes). Red, the sample (more ...)
Since the first human protein methyltransferase SUV39H1 was reported, tremendous efforts have been made in defining new PMTs and their biological functions.25,26 While such studies led to understanding biochemical functions of PMTs in vitro, the precise description of genome-wide chromatin-modifying sites of individual PMTs is lacking.4 Here we have addressed such need via reconstructing bioorthogonal posttranslation apparatus within living cells with engineered MAT and PMTs. The engineered reactomes can efficiently process membrane-permeable methionine analogues into SAM analogues within living cells and then label chromatins in situ. The advantage of developing such bioorthogonal system is to harness the specificity of genetic technique in generating a PMT variant that would only act on SAM analogue but native SAM, thus providing a means for functional analysis of a specific PMT in cells. While bioorthogonal approach has been successfully applied to other classes of enzymes, here we have demonstrated that multiple enzymes in common biochemical pathway can be engineered to sequentially generate and utilize cofactor analogues from corresponding precursor analogues in living cells.
The terminal-alkyne handle of the SAM analogues further allowed the modified chromatins to be enriched with the azido-containing biotin probe and then mapped with DNA sequencing technologies (CliEn-seq or CliEn-qPCR). In comparison with conventional ChIP-seq (chromatin immunoprecipitation with parallel DNA sequencing) experiments27, which cannot distinguish chromatin-modifying activities of closely-related enzymes (e.g. all the H3K9 PMTs), the merit of CliEn-seq lies in its ability to trace chromatin-modifying activities of a designated single PMTs in an unambiguous manner. Here we demonstrated one application of CliEn-seq by dissecting the activities of closely-related G9a and GLP1 (and likely PRMT1 as well) and shown that the two PMTs have complimentary H3K9 methylation activities across TSS and TES of target genes in HEK293T cells.
Extensive structural and biochemical investigations have led to the identification of conserved structural motifs of PMTs and suggest a common mechanism for cofactor recognition and catalysis. Given that multiple PMTs can be engineered to process bulky SAM analogues for target labeling,13,17 the corresponding engineered apparatuses are expected to be readily coupled with the MAT2A variant. The resultant constructed posttranslation apparatus can be applied to map genome-wide chromatin-modifying sites within living cells. More importantly, the promiscuous MATs variants and the corresponding methionine analogues developed here can be coupled with other methyltransferases for target labeling.28 The present work thus documents a powerful approach and the corresponding reagents to reveal downstream readouts of diverse PMTs.
Synthesis, purification and characterization of S-(Pent-2-en-4-yny)-L-homocysteine (EnYn-Met, 1b), S-(Hex-2-en-5-yny)-L-homocysteine trifluoroacetic salt (Hey-Met, 1c), and S-[4-(prop-2-ynyloxy)but-2-enylthio]-L-homocysteine (Pob-Met)
S-Benzyl-L-homocysteine (255 mg, 1 mmol) was placed in a round-bottom flask connected to an ammonia cylinder. Into the flask with a dry ice-ethanol bath was added liquid ammonia (20 mL). Sodium metal (50 mg, 2.2 mmol) was then added portionwise (from a clear solution to a dark blue solution). Within few minutes, the solution became colorless. The dry ice-ethanol bath was removed to allow the evaporation of ammonia. The rest trace of ammonia was removed with the aid of a flow of argon, followed by vacuum. The resulting white solid was dissolved in dry ethanol (10 mL) and cooled to 0 °C. A solution of (E)-6-bromohex-4-en-1-yne, (E)-pent-2-en-4-ynyl tosylate, or (E)-1-bromo-4-(prop-2-ynyloxy)but-2-ene (1.05 mmol) in ethanol (3 mL) was added (see Supporting Information for the synthesis of these precursors). This reaction mixture was stirred at 0 °C for 1 h, followed by overnight at ambient temperature (22 °C). The ethanol was removed under reduced pressure. The resultant mixture was re-dissolved in distilled de-ionized (DD) water (5 mL) and passed through a Dowex 50® (H+) ion-exchange column. The column was washed with DD water until neutral, and then eluted with diluted ammonia hydroxide (5% DD water solution). The portions containing the expected products (monitored by silica gel TLC eluted with 4:1:1 v/v/v n-BuOH/AcOH/H2O and detected by 0.2% nihydrin in 16:3:1 v/v/v 96% EtOH/AcOH/s-collidine upon heating) were combined and concentrated on a rotary evaporator to yield white solids. The crude products were then dissolved in 0.1 N HCl and subject to the further purification with the preparative reversed-phase HPLC (XBridge™ Prep C18 5μm OBD™ 19×150mm). The final products were eluted with linear gradients from 10% to 60 % of acetonitrile containing 0.1% trifluoroacetic acid in aqueous trifluoroacetic acid (0.1%) in 15 min with a flow of 10 mL/min. The desired fractions were collected and lyophilized to yield white powders (yields of 56%, 35% and 58% of 1b, 1c and Pob-Met, respectively).
1H NMR for EnYn-Met 1b (500MHz, D2O+formic acid-d2): δ 1.98-2.04(m, 1H), 2.06-2.12(m, 1H), 2.50(t, 2H, J = 7.4Hz), 3.12(d, 2H, J = 7.4Hz), 3.15(s, 1H), 3.99(t, 1H, J = 6.0Hz), 5.50(d, 1H, 15.7Hz), 6.06-6.12(m, 1H);13C NMR (125MHz, D2O+formic acid-d2): δ 25.99, 30.15, 33.06, 52.86, 79.09, 82.71, 111.45, 142.06, 172.77; MS(ESI) m/z: 200 [M+H]+; HRMS: calculated for C9H14NO2S ([M+H]+) 200.0745, found 200.0746.
1H NMR for Hey-Met 1c (500MHz, D2O): 2.13-2.19(m, 1H), 2.22-2.28(m, 1H), 2.52(t, 1H, J = 2.6Hz), 2.66(t, 2H, J = 7.5Hz), 3.00-3.02(m, 2H), 3.23(dd, 2H, J = 7.2, 0.7Hz), 4.14(t, 1H, J = 6.3Hz), 5.64-5.70(m, 1H), 5.75-5.81(m, 1H);13C NMR (150MHz, D2O+formic acid-d2): δ 21.20, 25.68, 30.14, 32.68, 52.78, 71.94, 83.23, 117.00(q, J =289.78), 127.90, 128.37, 163.65(q, J =35.2Hz), 172.76; MS(ESI) m/z: 214 [M+H]+; HRMS: calculated for C10H16NO2S ([M+H]+) 214.0902, found 214.0898.
1H NMR for Pob-Met (500MHz, DMSO-d6): 1.74-1.82(m, 1H), 1.93-1.99(m, 1H), 2.54(t, 1H, J = 7.6Hz), 3.15(d, 2H, J = 5.8Hz), 3.20-3.33(m, 1H), 3.45(d, 1H, J = 2.4Hz), 3.98(d, 2H, J = 4.3Hz), 4.12(d, 2H, J = 2.4Hz),5.64-5.67(m, 2H), 7.54(brs, 2H); 13C NMR (150MHz, DMSO-d6): δ26.62, 31.06, 32.04, 53.14, 56.53, 68.79, 77.27, 80.32, 128.38, 129.49, 169.15; MS(ESI) m/z: 244 [M+H]+; HRMS: calculated for C11H17NO3NaS ([M+Na]+) 266.0827, found 266.0822.
Bacterial expression and purification of proteins
Full-length human MAT2A plasmid was provided by Dr. Udo Oppermann at University of Oxford. The N-terminal His6-tagged human MAT2A in a pNIC28-Bsa4 kanamysin-resistant vector ( was expressed in E. coli (DE3) Rosseta 2 strain. A single colony was inoculated into 5 mL Miller Luria-Bertani (LB) medium containing 50 μg/ml kanamycin and 34 μg/ml chloramphenicol and grown at 37°C overnight. 10 ml of the o/n culture was added into 1L LB medium and the resultant medium was allowed to grow at 37°C until OD600=0.6–0.8. The cells were then induced with 0.5 mM IPTG at 17 °C for another 16 h before harvesting. The cell pallets were collected and homogenized by French press Cell Disruptor (Thermo) in lysis buffer (50 mM Tris-HCl, pH=8.0), 50 mM NaCl, 5 mM β-mercaptoethanol, 25 mM imidazole, Roche protease inhibitor and 5% (v/v) glycerol. After centrifuging at 15,000 g for 1 h at 4 °C, MAT2A was purified by Ni-NTA agarose resin (Qiagen) according to the previous method17 followed by a 5-ml HiTrap-Q Sepharose XL column (GE healthcare) using a linear gradient from 35 mM to 1 M KCl in a buffer containing 25 mM Tris-HCl (pH=8.0) and 5% (v/v) glycerol. Fractions containing MAT2A were combined and concentrated using Amicon Ultra-10K centrifugal filter device. The protein concentrations were determined with Bradford assay kit (Bio-Rad) with BSA as a standard. The concentrated proteins were stored at −80°C prior to use. All MAT2A mutants were generated by using QuickChange site-directed mutagenesis kit (Stratagene) according the vendor’s protocols. The resulted mutant plasmids were confirmed by DNA sequencing. The mutants were expressed and purified as described above for the native protein. The G9a/GLP1 mutants and 5′-methylthioadenosine nucleosidase (MTAN) were expressed and purified as described previously.3-5
Mass spectrometric assays of in vitro coupled reactions of MAT2A and PMT variants
For the reactions using histone peptide as the substrate, 50 μM H3 (1-21) peptide was added into 100 μL mixture containing 50 mM Tris-HCl (pH 8.5), 100 mM KCl, 2 mM MgCl2, 2.5 mM ATP, 2.5 mM a methionine analogue, 7.5 μM MAT2A mutant, 1 μM G9a or GLP1 mutant, and 20 nM MTAN. The resultant reaction mixture was incubated for 10 h at ambient temperature (22 °C). For the reactions using RGG peptide as the substrate, 50 μM RGG peptide substrate was added into 100 μL mixture containing 50 mM Tris-HCl (pH=8.5), 100 mM KCl, 2 mM MgCl2, 2.5 mM ATP, 2.5 mM a methionine analogues, 7.5 μM MAT2A mutant, 2 μM PRMT1 mutant, and 20 nM MTAN. The resultant reaction mixture was incubated for 10 h at ambient temperature (22 °C). The samples were then concentrated to around 20 μL by a SpeedVac and injected into LC-MS for analysis according to the previously-described protocols.13,17,29
For the reactions with histone substrate, 10 μM recombinant human histone H3 (New England Biolab) was added into 100 μL mixture containing 50 mM Tris-HCl (pH=8.5), 100 mM KCl, 2 mM MgCl2, 2.5 mM ATP, 2.5 mM a methionine analogue, 7.5 μM MAT2A mutant, 1 μM G9a or GLP1 mutant, and 20 nM MTAN. The resultant reaction mixture was incubated for 24 h at ambient temperature (22°C). The histone-containing reaction samples were passed through a Sep-Pak Vac C18 1cc column (Waters, Milford, MA). The protein mixture was then eluted with 1:1 acetonitrile and 0.1% TFA, concentrated with a SpeedVac and resolved by SDS-PAGE (Criterion™ Precast gel, 18% Tris-HCl, Bio-Rad) followed by staining with Coomassie Blue (Bio-Rad). The corresponding histone H3 gel bands were subject to LC/MS/MS analysis according to the method described previously.13,17,29
Cloning and mutagenesis of mammalian vectors
Full-length N-terminal flag-tagged human MAT2A was constructed by PCR using pNIC28-Bsa4-MAT2A (provided by Dr. Udo Oppermann at University of Oxford) as the template. The following primers were used: 5′-GATTAGAATTCATGGACTACAAAGACGATGACGAT AAAAACGGACAGCTCAACGGCTTCCAT-3′ (forward with the Flag tag sequence in bold); 5′-TAATCCTCGAGTCAATATTTAAGCTTTTTGGGCACTT CCCATGGGAAGCTGTC-3′ (reverse). The resultant PCR product was cloned into pcDNA3 vector through ECORI/XhoI restrictive sites to yield the pcDNA3-Flag-MAT2A plasmid. The I117A variant of pcDNA3-Flag-MAT2A plasmid was generated with the pcDNA3-Flag-MAT2A plasmid as the template with Quick change site-directed mutagenesis kit (Stratagene) according to the manufacturer’s instruction. The plasmids of pcDNA3-Flag-G9a/GLP1 and their variants were constructed as described previously.29
Transfection of G9a/GLP1/MAT2A variants followed by the treatment of methionine analogues
HEK293T cells (ATCC) were cultured in Dulbecco’s Modified Eagle Medium (DMEM) with 10% (v/v) Fetal Bovine Serum (FBS) and maintained in a humidified incubator at 37 °C and 5 % CO2. The cells were seeded into CellBIND surface T75 flask (Corning) and transiently transfected with G9a/GLP1/MAT2A plasmids (native or mutants) in Lipofectamine 2000 (Invitrogen) at 30–40% confluence according to the manufacturer’s instructions. The cells were washed with PBS 12 h after transfection and recovered in fresh DMEM medium for 24 h. The medium was then removed and the cells were incubated in methionine-deficient DMEM medium (Invitrogen) supplemented with 10 % dialyzed FBS (Gibco) for 30 min in the CO2 incubator to deplete the intracellular methionine. The medium was then replaced with the methionine-deficient DMEM medium supplied with 10% dialyzed FBS and various concentrations of methionine analogues. The cells were incubated for 24 h unless mentioned otherwise and then harvested. The tranfection efficiency of the plasmids was confirmed by western blotting using the following antibodies: anti-Flag (Sigma, Cat. No: F3165); anti-G9a (Millipore, Cat. No: 07-551); anti-GLP1 (Millipore, Cat. No: 09-078; anti-MAT2A (Abcam, Cat. No: ab77471). The Western blotting was performed as described before.29
Measurement of the production of Hey-SAM (2c) within living cells
HEK293T cells were transfected with mock (empty pcDNA3 vector) or the MAT2A I117A plasmid and then incubated with different concentrations of Hey-Met 1c according to the protocols described above. After 24-h incubation, the cells were harvested and lysed in ice-cold RIPA buffer (Sigma) containing 1×Complete EDTA-free protease inhibitor cocktail (Roche). The cell lysates was then mixed with equal volume of 20 % HClO4. The samples were passed a 3K centrifugal filter (Millipore) to remove the precipitated materials. The resultant flow-through was collected and resolved by reversed-phase HPLC (DELTA PAK C18 15μm, 300×3.9 mm) by monitoring UV260 nm with a gradient of acetonitrile in aqueous trifluoroacetic acid (0.01%) from 0% to 10% in 15 min and then to 70% in 5 min at a flow rate of 1 mL/min. The resultant Hey-SAM 2c was collected and confirmed by ESI-MS (Figures 3b, S5). To further quantify the amount of Hey-SAM 2c generated within living cells, about 107 HEK293T cells that were transfected with the MAT2A I117A and then treated with 1mM Hey-Met 1c were lysed as described above. After removing the precipitated materials, the known amount of 5′-deoxy-methylthioadenosine (MTA) was added as the internal standard before injecting to HPLC. The peak area of Hey-SAM was integrated and compared to that of MTA to quantify the amount of Hey-SAM produced within the cells.
Alamar blue assay for cell viability
HEK293T cells (with or without transfection of MAT2A I117A plasmid) were resuspended in DMEM medium containing 10% FBS, and then seeded into 96-well plates (100μL/well) at concentration of 105 cells/mL for an initial 12 h period for cell attachment. Subsequently, the medium was replaced with 100 μL methionine-deficient DMEM medium containing the varied concentrations of Hey-Met 1c. After incubating with the compound for 8 h, cell viability was determined by adding Alamar blue (Invitrogen, Cat. No: DAL1100) into the culture medium at the final concentration of 10 % and the plate was returned to the incubator. Optical density of the plate was measured at the excitation wavelength of 530 nm and the emission wavelength at 590nm with a standard spectrophotometer 4 h after adding Alamar blue. The results were presented as a percentage of the control without adding Hey-Met 1c.
In-gel fluorescence of proteome-wide substrates of G9a within living cells
HEK293T cells were transfected with the engineered MAT2A and/or G9a mutant plasmids and then treated with 1 mM Hey-Met 1c for a varied period of time. The resultant cells were harvested and lysed in an ice-cold RIPA buffer containing 1×Complete EDTA-free protease inhibitor cocktail. After centrifuge, the supernatant was passed through the detergent removal spin column (Pierce, Cat. No: 87778) and eluted with Tris buffer (100 mM Tris, pH=8.0, 10% glycrol, 1mM TCEP). 50μg cell lysates (protein concentration was determined by Bradford assay, Bio-Rad Laboratories) were subjected to CuAAC, followed by in-gel fluorescence as described previously.30
Pull-down and acid extraction of histones in cellular contexts
HEK293T cells were transfected with the engineered MAT2A and G9a or GLP1 mutant plasmids and then treated with 1 mM Hey-Met (1c) as described above. The resultant cells were harvested and lysed in ice-cold RIPA buffer containing 1×Complete EDTA-free protease inhibitor cocktail. Into 2 mg of the lysates was added the CuAAC ligation cocktail (1 mM CuSO4, 100 μM Tris [(1-benzyl-1H-1,2,3-triazol-4-yl) methyl] amine (TBTA), 2.5 mM Tris (2-carboxyethyl) phosphine hydrochloride (TCEP), and 200 μM azido-diazo-biotin as the final concentrations).29,31 The mixture was further incubated for 1 h at ambient temperature (22 °C).31 The reaction was quenched by the addition of 10 mL cold methanol into each sample and kept at −80 °C overnight. Precipitated proteins were centrifuged at 3,000 rpm for 30 min, washed twice with 10 mL cold methanol, and dried for 20 min. The tried protein samples were redissolved in 400 μL dilution buffer (50 mM triethylamine at pH=7.4, 150 mM NaCl, 1×Complete EDTA-free protease inhibitor cocktail, and 0.5% SDS) with gentle vortexing and sonication. For each sample, 50 μg protein was spared as loading control of western blotting, and the remaining samples were mixed with 50 μL Streptavidin Sepharose High Performance beads (GE Healthcare) that were pre-washed and resuspended in the dilution buffer, and rotated end-to-end at ambient temperature (22°C) for 1 h. The samples were then washed once with PBS containing 0.2% SDS, twice with PBS, and once with 250 mM ammonium bicarbonate buffer. The proteins were finally eluted for 30 min with 500 μL elution buffer containing 25mM sodium dithionite and 250 mM ammonium bicarbonate at ambient temperature (22 °C) and concentrated by Amicon Ultra centrifugal filters (Millipore). The eluted protein samples were then snap-frozen, lyophilized into powder, and dissolved in 1×loading buffer (Bio-rad). These samples were resolved by SDS-PAGE and analyzed by western blotting using anti-histone H3 antibody (Millipore, Cat. No: 05-928) according to the previous protocol.29 To extract core histones, HEK293T cells were first resuspended in 1×hypotonic lysis buffer (Sigma, Cat. No: L9161) supplemented with 1×Complete EDTA-free protease inhibitor cocktail and incubated on ice for 15 min. In the resultant cell lysates was added NP-40 to a final concentration of 0.6 %, followed by vigorous vortexing of the samples for 10 s. The samples were then centrifuged at 11,000 g at 4°C for 10 min. The nuclear pellet was washed once in the ice-cold hypotonic buffer, and resuspended in 0.4 N H2SO4 and extracted at 4 °C overnight. After centrifuging the samples at 11,000g at 4°C for 10 min, the supernatant containing core histones were collected and resolved in 18 % SDS-PAGE. The corresponding gel bands were excised and subject to LC/MS/MS analysis as described previously.13,17,29
Native chromatin immunoprecipitation (ChIP) and Clickable Chromatin Enrichment (CliEn)
For Native chromatin immunoprecipitation (ChIP), the nuclei were isolated from 107 HET293T cells using 1×hypotonic lysis buffer (Sigma) as described above, re-suspended in the digestion buffer (0.32 M sucrose, 50 mM Tris-HCl at pH=7.5, 4 mM MgCl2, 1 mM CaCl2, 0.1 mM PMSF, and 1×Complete EDTA-free protease inhibitor cocktail), and then digested with micrococcal nuclease (MNase, Sigma, Cat. No: N3755). Here MNase’s concentrations and digestion time were optimized for each set of experiments according to the reported protocol.32 The MNase-mediated digestion was carried out for 10 min at 37 °C and stopped by the addition of EDTA to the final concentration of 5 mM. The resultant samples were then centrifuged at 10,000 g for 5 min, and the supernatant was collected as S1 fraction. The remaining pellet was resuspended in 1 mL lysis buffer (1 mM Tris-HCl, pH=7.4, 0.2 mM EDTA, 0.2 mM PMSF, and 1×Complete EDTA-free protease inhibitor cocktail) and dialyzed against the same buffer using 2K MWCO Slide-A-Lyzer Dialysis Cassette (Thermo Scientific) for 3 h at 4 °C. After dialysis, the samples were centrifuged at 10,000 g for 5 min, and the supernatant was collected as S2 fraction. The soluble chromatin was prepared by combining S1 and S2 fractions. With the samples, immunoprecipitation was carried out with anti-H3K9me2 antibody (Millipore, Cat. No: 17-681) coupled with 100 μL protein G magnetic beads (Millipore) at 4°C overnight. The beads enriched with the target proteins were then washed sequentially with Buffer A (50 mM Tris-HCl at pH=7.5, 10mM EDTA and 50 mM NaCl), Buffer B (50 mM Tris-HCl at pH=7.5, 10 mM EDTA, and 100 mM NaCl) and Buffer C (50 mM Tris-HCl at pH=7.5, 10 mM EDTA, and 150 mM NaCl). The bound chromatin was eluted with 500 μL elution buffer (20 mM Tris-HCl at pH=7.5, 50 mM NaCl, 1×Complete EDTA-free protease inhibitor cocktail, and 1% SDS) with agitation for 30 min at ambient temperature (22 °C). The chromatin samples were digested with Proteinase K (Sigma, Cat. No: P2308) at 55 °C for 2 h and the DNA was purified by QIAquick PCR Purification columns (Qiagen) using Buffer PB, and analyzed by quantitative PCR (qPCR) as described below.
For Clickable Chromatin Enrichment (CliEn), the soluble chromatins, prepared above prior to the step of antibody immunoprecipitation, were subject to the CuAAC cocktail (1 mM CuSO4, 100 μM Tris [(1-benzyl-1H-1,2,3-triazol-4-yl) methyl] amine (TBTA), 2.5 mM Tris (2-carboxyethyl) phosphine hydrochloride (TCEP), and 200 μM azido-diazo-biotin as the final concentrations) for 1 h at ambient temperature (22 °C) with rotation. Into the mixture was added 100 μL streptavidin beads, which were pre-washed and resuspended in the incubation buffer (20 mM Tris-HCl at pH=7.5, 50 mM NaCl and 1×Complete EDTA-free protease inhibitor cocktail). The resultant sample was mixed gentlely with a rotator for another 3 h at 4 °C. The streptavidin beads were then sequentially washed with 10 ml PBS containing 0.1% SDS, the 2× incubation buffer (20 mM Tris-HCl at pH=7.5, 50 mM NaCl and 1×Complete EDTA-free protease inhibitor cocktail). The chromatins were eluted with 500 μL elution buffer (25 mM sodium dithionite, 1% SDS, 20mM Tris-HCl, pH=7.5, 50mM NaCl, and 1×Complete EDTA-free protease inhibitor cocktail) with agitation for 30 min at ambient temperature (22 °C). The eluates were digested with Proteinase K at 55 °C for 2 h and the DNA was purified by QIAquick PCR Purification columns (Qiagen) using Buffer PB, and analyzed by quantitative PCR (qPCR) as described below. The bound and input DNA samples from native chromatin immunoprecipitation (ChIP) or Clickable Chromatin Enrichment (CliEn) were analyzed by qPCR with SYBR Green PCR Master Mix (Applied Biosystems) using the ABI Prism 7500 fast real-time PCR machine. Primers for qPCR were provided in Table S1. Relative ChIP-qPCR values were presented as the fold of changes versus the control groups (without transfection of native G9a or GLP1); CliEn-qPCR data were presented similarly as the fold of changes versus the control groups (without transfection of MAT2A I117A, G9a Y1154A or GLP1 Y1211A variants).
Samples and data process for CliEn-seq
CliEn samples were prepared for sequencing using Illumina TruSeq DNA Sample Preparation Kit according to the standard preparation protocol ( with some modifications. The CliEn DNA and input DNA concentrations were first measured by Bioanalyzer (Agilent) (for cells tranfected with MAT2A I117A: CliEn DNA 2.44 ng, input DNA 2078.75 ng; for cells co-transfected with MAT2A I117A and G9a Y1154A: CliEn DNA 115.44 ng, input DNA 2058.70 ng; for cells co-transfected with MAT2A I117A and GLP1 Y1211A: CliEn DNA 84.61 ng, input DNA 2305.15 ng) and presented in Figure 4b. Next, ~5ng of the CliEn DNA or input DNA was used for library preparation and a 1:10 dilution of the DNA adapters were ligated. DNA fragments were then amplified with PCR Primer Cocktail (6 cycles) and 1/10 of the PCR product was checked on the Agilent Bio-analyzer. The purified DNA libraries were diluted to 10 nM for loading on the flow cell. Sequencing service was performed on an Illumina Hiseq 2000 sequencer according to the standard Illumina protocol resulting in ~ 15 million quality-filtered sequences of length 75/75 per sample. CliEn-seq reads were aligned to the reference human genome (hg19, NCBI Build 37) using the BWA program33 and the reads mapped to multiple locations were filtered out with only one position saved. The sequence results were processed, normalized by total reads, and displayed using the GenPlay tool. The Illumina HiSeq data have been deposited into GEO (Accession Number GSE42792). To generate the CliEn-seq read density profiles, we used the ChIPseeqer-2.0.34 A read density matrix was generated with the average number of reads in 10 bp bins for a windown of +/- 2 kb around the transcriptionn start site (TSS) or the transcriptionn end site (TES) of the 35 selected genes (Table S2). The read density plot was created with the mean value per bin in the matrix.
Supplementary Material
We thank Drs. Min, Huang, Oppermann and Schramm for plasmids; Genomics Core Laboratory and Dr. Agnes Viale at MSKCC for DNA sequencing service; NIGMS (1R01GM096056), the NIH Director’s New Innovator Award Program (1DP2-OD007335), the V Foundation for Cancer Research, March of Dimes Foundation, Starr Cancer Consortium, and the Alfred W. Bressler Scholars Endowment Fund for financial supports.
Supporting Information. Additional procedures, synthetic methods and characterizations for the precursors of methionine analogues, NMR spectra and supplementary figures. This material is available free of charge via the Internet at
Data Deposition. The Illumina HiSeq data were deposited into GEO with Accession Number GSE42792.
1. Schubert HL, Blumenthal RM, Cheng X. Trends Biochem Sci. 2003;28:329. [PMC free article] [PubMed]
2. Greer EL, Shi Y. Nat Rev Genet. 2012;13:343. [PubMed]
3. Luo M. ACS Chem Biol. 2012;7:443. [PMC free article] [PubMed]
4. Geiman TM, Robertson KD. J Cell Biochem. 2002;87:117. [PubMed]
5. Clapier CR, Cairns BR. Annu Rev Biochem. 2009;78:273. [PubMed]
6. Fritsch L, Robin P, Mathieu JR, Souidi M, Hinaux H, Rougeulle C, Harel-Bellan A, Ameyar-Zazoua M, Ait-Si-Ali S. Mol Cell. 2010;37:46. [PubMed]
7. Garcia-Cao M, O’Sullivan R, Peters AH, Jenuwein T, Blasco MA. Nat Genet. 2004;36:94. [PubMed]
8. Kim KC, Geng L, Huang S. Cancer Res. 2003;63:7619. [PubMed]
9. Xu PF, Zhu KY, Jin Y, Chen Y, Sun XJ, Deng M, Chen SJ, Chen Z, Liu TX. Proc Natl Acad Sci U S A. 2010;107:2521. [PubMed]
10. Komoto J, Yamada T, Takata Y, Markham GD, Takusagawa F. Biochemistry. 2004;43:1821. [PubMed]
11. Lin Q, Jiang F, Schultz PG, Gray NS. J Am Chem Soc. 2001;123:11608. [PubMed]
12. Shinkai Y, Tachibana M. Genes Dev. 2011;25:781. [PubMed]
13. Islam K, Zheng W, Yu H, Deng H, Luo M. ACS Chem Biol. 2011;6:679. [PMC free article] [PubMed]
14. Cantoni G. J Am Chem Soc. 1952;74:2942.
15. Catoni GL. J Biol Chem. 1953;204:403. [PubMed]
16. Bedford MT, Clarke SG. Mol Cell. 2009;33:1. [PMC free article] [PubMed]
17. Wang R, Zheng W, Yu H, Deng H, Luo M. J Am Chem Soc. 2011;133:7648. [PMC free article] [PubMed]
18. Melnyk S, Pogribna M, Pogribny IP, Yi P, James S. J Clin Chem. 2000;46:265. [PubMed]
19. Yang YY, Ascano JM, Hang HC. J Am Chem Soc. 2010;132:3640. [PMC free article] [PubMed]
20. Couture JF, Collazo E, Ortiz-Tello PA, Brunzelle JS, Trievel RC. Nat Struct Mol Biol. 2007;14:689. [PubMed]
21. Kruidenier L, Chung CW, Cheng Z, Liddle J, Che K, Joberty G, Bantscheff M, Bountra C, Bridges A, Diallo H, Eberhard D, Hutchinson S, Jones E, Katso R, Leveridge M, Mander PK, Mosley J, Ramirez-Molina C, Rowland P, Schofield CJ, Sheppard RJ, Smith JE, Swales C, Tanner R, Thomas P, Tumber A, Drewes G, Oppermann U, Patel DJ, Lee K, Wilson DM. Nature. 2012;488:404. [PubMed]
22. Ling BM, Bharathy N, Chung TK, Kok WK, Li S, Tan YH, Rao VK, Gopinadhan S, Sartorelli V, Walsh MJ, Taneja R. Proc Natl Acad Sci U S A. 2012;109:841. [PubMed]
23. Kubicek S, O’Sullivan RJ, August EM, Hickey ER, Zhang Q, Teodoro ML, Rea S, Mechtler K, Kowalski JA, Homon CA, Kelly TA, Jenuwein T. Mol Cell. 2007;25:473. [PubMed]
24. Rathert P, Dhayalan A, Murakami M, Zhang X, Tamas R, Jurkowska R, Komatsu Y, Shinkai Y, Cheng X, Jeltsch A. Nat Chem Biol. 2008;4:344. [PMC free article] [PubMed]
25. Rea S, Eisenhaber F, O’Carroll D, Strahl BD, Sun ZW, Schmid M, Opravil S, Mechtler K, Ponting CP, Allis CD, Jenuwein T. Nature. 2000;406:593. [PubMed]
26. Martin C, Zhang Y. Nat Rev Mol Cell Biol. 2005;6:838. [PubMed]
27. Collas P. Mol Biotechnol. 2010;45:87. [PubMed]
28. Dalhoff C, Lukinavicius G, Klimasauskas S, Weinhold E. Nat Chem Biol. 2006;2:31. [PubMed]
29. Islam K, Bothwell I, Chen Y, Sengelaub C, Wang R, Deng H, Luo M. J Am Chem Soc. 2012;134:5909. [PMC free article] [PubMed]
30. Bothwell IR, Islam K, Chen Y, Zheng W, Blum G, Deng H, Luo M. J Am Chem Soc. 2012;134:14905. [PMC free article] [PubMed]
31. Yang YY, Grammel M, Raghavan AS, Charron G, Hang HC. Chem Biol. 2010;17:1212. [PMC free article] [PubMed]
32. Wagschal A, Delaval K, Pannetier M, Arnaud P, Feil R. CSH Protoc. 2007 2007, pdb prot4767. [PubMed]
33. Li H, Durbin R. Bioinformatics. 2009;25:1754. [PMC free article] [PubMed]
34. Giannopoulou EG, Elemento O. BMC Bioinformatics. 2011;12:277. [PMC free article] [PubMed]