|Home | About | Journals | Submit | Contact Us | Français|
Protein methyltransferases (PMTs) utilize S-adenosyl-L-methionine as a cofactor and deliver its sulfonium methyl moiety to diverse substrates. These methylation events can lead to meaningful biological outcomes from transcriptional activation/silencing to cell cycle regulation. With the long-term goal of elucidating the substrates and defining the functions of PMTs, our laboratory recently developed technology based on protein engineering in tandem with SAM analogue cofactors and bioorthogonal click chemistry to unambiguously profile the substrates of a specific PMT. The following protocols encapsulate the logic and methods of selectively profiling the substrates of a candidate PMT by (1) engineering the selected PMT to accommodate a bulky SAM analogue; (2) generating the proteome containing the engineered PMT; (3) visualizing the proteome-wide substrates of the designated PMT via bioorthogonal labeling with a fluorescent tag and finally (4) pulling down the proteome-wide substrates for mass spectrometric analysis.
To maintain genome stability, human DNA is wrapped around the core histone octamer (H2A, H2B, H3 and H4) to form nucleosomes. These basic structural repeats can be further organized into the higher order architecture of chromatins and eventually chromosomes (Qiu, 2006). To program the chromatins for gene expression or repression, the histone tails undergo frequent posttranslational modifications, such as methylation, acetylation, phosphorylation, and ubiquitylation. The combinations of these modifications have been shown to modulate downstream signals and thus render meaningful biological outcomes (Kouzarides, 2007).
Histone methylation is catalyzed by more than 60 human protein methyltransferases (PMTs), which use S-adenosyl-L-methionine (SAM) (Fig. 1b) as a cofactor and deliver SAM’s sulfonium methyl group to specific arginine or lysine residues on histones (Kouzarides, 2007). The site and extent (mono-, di- or tri-) of histone methylation are tightly associated with the biological readouts of gene transcription or silencing (Martin and Zhang, 2005). For instance, the tri-methylation of histone 3 lysine 4 (H3K4) by multiple PMTs (e.g. MLL1–5 and SET1A/B) is linked to transcriptional activation, while the di/tri-methylation of histone 3 lysine 9 (H3K9) by G9a, GLP and SUV39H1/2 correspond to gene silencing (Lachner and Jenuwein, 2002; Martin and Zhang, 2005). Among 9 human protein arginine methyltransferases (PRMTs), PRMT1, CARM1/PRMT4, PRMT5 and PRMT6 have demonstrated in vivo methylation activities on histone substrates (Bedford and Clarke, 2009).
Accumulated evidence shows that PMTs also act on diverse non-histone targets. Among such examples are the methylation of TAF10, ERα, PCAF, NF-κB, DNMT1, HIV transactivator Tat and FOXO3 by SET7/9 (Calnan et al., 2012; Ea and Baltimore, 2009; Esteve et al., 2009; Kouskouti et al., 2004; Masatsugu and Yamamoto, 2009; Mowen et al., 2001; Pagans et al., 2010; Subramanian et al., 2008; Yamagata et al., 2008; Zhao et al., 2008), transcription factors STAT1, RUNX1, and FOXO1 by PRMT1, and transcription coactivators p300 and CBP by CARM1 (Chevillard-Briet et al., 2002; Xu et al., 2001). Certain non-histone targets can be methylated by multiple PMTs (e.g. the tumor suppressor p53 by SET7/9, SET8, SMYD2, G9a, and GLP) (Chuikov et al., 2004; Huang et al., 2010; Huang et al., 2006; Shi et al., 2007). The collective actions of protein-methylating enzymes, or ‘writers’ of methylation, orchestrate physiological functions of the substrates in an epigenetic manner (Strahl and Allis, 2000). Aberrant methylation activities have been implicated in many human diseases including cancer (Bhaumik et al., 2007; Chi et al., 2010).
Elucidating methylation landscape (or methylome) of an individual PMT is essential in order to elucidate its physiological or pathological roles. To address the need, our laboratory recently formulated the concept of Bioorthogonal Profiling of Protein Methylation (BPPM). Here specific PMTs can be engineered to accommodate the SAM analogues and thus transfer a bioorthogonal azido or alkynyl moiety to respective substrates (Fig. 1a) (Islam et al., 2012; Islam et al., 2011; Wang et al., 2011b). Given that the synthetic SAM analogues are bulkier than native SAM, they are expected to be inert toward native PMTs and thus selectively label the substrates of designated (engineered) PMTs. In the previous work, we have demonstrated that the SAM-binding pockets of G9a and GLP can be tailored via site-directed mutagenesis. The exogenously expressed mutant proteins were screened against 4-azidobut-2-enyl SAM (Ab-SAM). The matched mutant-cofactor pairs were identified to modify the substrates of G9a and GLP. The resultant azido-labeled proteins were readily visualized with commercially available TAMRA-DIBO (“Click-iT”) fluorescent dye or subject to biotinylation with a “Click-iT” biotin probe for target pulldown and identification (Fig 1b) (Islam et al., 2012).
With G9a and Ab-SAM as a paradigm, the current protocols present the step-wise development of the BPPM methodology (Fig 1c). We first describe the procedure relying on matrix-assisted laser desorption/Ionization mass spectrometry (MALDI-MS) to identify an active PMT mutant toward Ab-SAM (BASIC PROTOCOL 1). We then describe the procedure of transiently expressing and harvesting full-length, active, PMT variant in its cellular context (BASIC PROTOCOL 2). The enzymatic transferring of the azido moiety from Ab-SAM was then carried out in the context of the cellular proteome containing the engineered PMT variant. The azido-labeled proteome was then visualized with TAMRA-DIBO via in-gel fluorescence (BASIC PROTOCOL 3). In parallel, the Ab-SAM-labeled cell lysates were treated with biotin-DIBO, followed by target pulldown with streptavidin beads and target identification with mass spectrometer (BASIC PROTOCOL 4). The overall approach is featured by its robustness and generality to profile the targets of designated PMTs via BPPM.
A set of bulky SAM derivatives containing terminal-alkyne and sulfonium-β-sp2 substitutions have been reported as suitable cofactor surrogates of native or engineered PMTs (e.g. (E)-pent-2-en-4-ynyl, (E)-hex-2-en-5-ynyl, 4-propargyloxy-but-2-enyl SAM for native MLL4, engineered G9a and PRMT1, respectively). We recently expanded the cofactor repertoire by including 4-azido-but-2-enyl S-adenosyl-L-methionine (Ab-SAM) (Islam et al., 2012). While these SAM analogues all contain the catalytically essential sulfonium-β-sp2 substitution and the terminal alkyne/azido groups for azide-alkyne cycloaddition (CuAAC or click chemistry), Ab-SAM solely stands out as a BPPM reagent on the basis of its distinct sulfonium-δ azido moiety, a labeling handle favored for less toxic copper-free click chemistry (Baskin et al., 2007). The distinct geometry, size and polarity of the azide group (a linear 3-atom dipolar) may also restrain this cofactor from being recognized by native PMTs and thus make it more suitable for BPPM even in the presence of irrelevant cellular PMTs. These features motivated us to use Ab-SAM as a demo to document the BPPM protocols for proteome-wide labeling of PMT substrates.
As described previously, our lab has been successful in engineering multiple PMTs to process SAM analogue cofactors for BPPM (Islam et al., 2012; Islam et al., 2011; Wang et al., 2011b). A common strategy in these cases is to generate and screen a collection of PMT mutants as BPPM candidates. Although the choices of the PMT mutants vary case by case, our previous experiences were to focus on tailoring the residues that are adjacent to the binding site of SAM’s methyl group according to available PMT structures. For the PMTs whose structures have not been reported, we relied on sequence alignment to identify the comparable SAM-recognizing residues. In BASIC PROTOCOL 1, G9a was chosen to exemplify the selection and screening process of identifying active PMT variants for BPPM. X-ray structures and sequence alignments of G9a with other SET-domain-containing PMTs revealed that a set of conserved aromatic residues are located at the interface between the channel for substrate lysine binding and the pocket for SAM interaction (Fig 2a, 2b). We anticipated that mutating the bulky hydrophobic residue(s) may vacant space in the active site allowing the accommodation of the bulky Ab-SAM analogue, while maintaining the transalkylation activity of native G9a (Islam et al., 2012), we carried out the screening with a single Ala mutation in this defined region (Y1067, Y1085, F1152, Y1154 and F1158 in Fig 2b). These residues directly contact with the substrate or SAM cofactor as revealed by G9a’s structures (Islam et al., 2012; Islam et al., 2011) and contain no posttranslational modification as reported so far. The corresponding G9a mutants were expressed in E. coli and purified with conventional methodologies (Islam et al., 2012). The G9a variants identified here for BPPM application are expected to be stable under the conditions of expression and purification, and thus enable the subsequent activity assay (STRATEGIC PLANNING below & BASIC PROTOCOL 1). We were able to identify the active G9a mutant Y1154A from the small, rational collection of G9a variants (BASIC PROTOCOL 1). As shown for the non-SET-domain arginine methyltransferase PRMT1 (Wang et al., 2011b), the concept of PMT engineering can be general for other PMTs and SAM analogues.
Many PMTs can recognize synthetic peptides as well as full-length proteins as substrates. Given the availability of synthetic peptides through solid-phase peptide synthesis and their ready characterization with mass spectrometer, we preferred peptides as substrates for initial in vitro screening of matched enzyme-cofactor pairs. G9a di-methylates a H3K9 peptide derived from its in vivo substrate histone H3 [ARTKQTARKSTGGKAPRKQLA(amino acids 1–21)] (Rathert et al., 2008). This peptide can be readily prepared with a solid-phase peptide synthesizer, purified to > 95% with HPLC and characterized by MALDI-TOF mass spectrometry. MALDI-TOF mass spectrometry was then applied to monitor the ability of the G9a mutants to modify the peptide substrate with Ab-SAM. The reaction with native G9a and SAM was carried in parallel as a positive control to avoid the false negatives, which may arise from the assay buffer or instrument failure. Our previous work has shown that H3K9 alkylation does not affect the ionization efficiency of the peptide and thus allows a direct quantification of the alkylation according to the peak ratios of mass spectrometric signals (Islam et al., 2011). Native G9a modifies the H3K9 21-aa peptide substrate with a high efficiency. If it were not the case, the optimization of peptide length (generally longer peptide) or the use of alternative peptide substrates should be considered at this stage. Prior to the assay, the reaction conditions with native PMTs and SAM must be optimized for buffer components (e.g. pH, detergents and salts), reaction time, and the concentrations of enzymes, peptide substrates and cofactors. If a large panel of enzyme-cofactor pairs (e.g. a 100 × 30 panel) needs to be screened and the capacity of MALDI-TOF is insufficient, an alternative high throughput, fluorogenic PMT assay can be used as detailed previously (Wang et al., 2011a).
For the proteome-wide BPPM with Ab-SAM, the engineered PMTs can either be exogenously expressed and then mixed with cell lysates or expressed within cells of interest via transient transfection. We noticed that the labeling efficiency of Ab-SAM often correlates to the amount of the exogenously-added or transiently-expressed PMT variants in the cell lysates. In addition, the constructs of PMT variants affect Ab-SAM’s efficiency to label targets. For instance, although the engineered methyltransferase catalytic domain of G9a is active in vitro, the substrate labeling of the G9a construct is barely detectable in the presence of complex cellular proteome. In contrast, the transiently-expressed full-length G9a robustly labels its substrates in the same cellular context (Islam et al., 2012). Although the cause of this difference remains to be explored, it is likely that other coactivators in the cellular context play the role to promote the activity of the full-length G9a variant via the interaction with G9a’s non-catalytic regions. In the current protocols, after identifying the active G9a variant for Ab-SAM, we relied on the transient transfection approach to overexpress the corresponding full-length G9a variant in HEK293T cells for BPPM.
To further enhance the substrate labeling efficiency of engineered PMTs and Ab-SAM, it will be desirable to inhibit the methylation activities of endogenous PMTs of interest resulting in a hypomethylated proteome as the substrate candidates for BPPM. To achieve this goal, the cells can be pre-treated with pan- or target-specific PMT inhibitors (Herrmann et al., 2005; Vedadi et al., 2011). In the current protocol with engineered G9a and Ab-SAM, HEK293T cells were treated with adenosine-2’,3’-dialdehyde (Adox, Sigma A7154) to generate the hypomethylated proteome. Adox is a potent inhibitor of S-adenosylhomocysteine hydrolase, which is responsible for the intracellular degradation of S-adenosylhomocysteine (SAH), a pan-PMT inhibitor. The Adox treatment elevates SAH’s intercellular concentration and thus reduces proteome-wide methylation (Herrmann et al., 2005).
The following is a standard protocol for the identification of pairs of active PMT mutants and SAM analogue cofactor, exemplified with Ab-SAM and G9a (aa 913 – 1193). Here the exogenously expressed and purified native G9a and G9a variants were incubated with the aa 1–21 H3K9 peptide and Ab-SAM. The extent of Ab-SAM-mediated transalkylation on the peptide was quantified by MALDI-MS. The synthesis of Ab-SAM (Fig 1c) is described in SUPPORTING PROTOCOL 4. This protocol is generally applicable to other PMTs and SAM analogue cofactors of interest.
The protein expression and purification were reported before and described as the following. Plasmids containing N-terminal His6-tagged methyltransferase SET domain of human G9a (aa 913−1193) were obtained from Dr. Jingrong Min’s lab at the University of Toronto. Mutants were generated by the QuickChange site-directed mutagenesis (Stratagene) following manufacturer’s protocol. The plasmid was transformed into E. coli Rosetta-2(DE3) strain (Novagen) using pET28a-LIC vector. Protein expression was induced in the presence of 25 µM ZnSO4 at 17 °C overnight with 0.5 mM IPTG. Protein was purified first using Ni-NTA agarose resin (Qiagen) and then by gel exclusion chromatography (Superdex-75, GE Healthcare). The concentrated protein was stored at −80 °C before use.
Keep all reagents on ice to avoid decomposition.
An example of expected results is shown in Fig. 3a, 3b.
The cells of interest were transiently transfected with the identified active G9a variant (BASIC PROTOCOL 1). The cells are then cultured in the presence of Adox to generate hypomethylation proteome, harvested and lysed for BPPM.
The optimal growth, transfection conditions (media type, confluency of cells, the amount of plasmids etc.) and the required amount of protein for BPPM analysis depend on the types of examined cells. These parameters need to be optimized prior to the following experiments. The parameters of HEK 293T cells used in BASIC PROTOCOL 3, 4 are provided here.
The size of cell culture flask is determined by the amount of protein required for BPPM. T-75 and T-25 flasks (353108/353136 - BD Falcon™ 25/75 cm2 Cell Culture Flask) are used for in-gel fluorescence imaging and proteomic pulldown, respectively (we routinely transfect mutants directly in flasks). The amounts of reagents such as Lipofectamine, Opti-MEM, or individual plasmids are proportional to the amount of treated cells. These parameters should be optimized prior to reach maximal transfection efficiency confirmed by Western blot with proper antibodies. The following protocol was described for a T-25 flask.
A media exchange step after 8 -12 h incubation is required for cells with high sensitivity towards Lipofectamin, although the step is not necessary for HEK293T cells.
Use of trypsin to harvest cells at this step is not recommended because the treatment may degrade cell surface proteins.
The purpose of the step is to remove serum proteins.
The cell pellets at this stage can be stored as the pellet at −80 °C for < 2 months. Allocate the samples for Western blot to confirm transfection efficiency and BASIC PROTOCOL 3 and 4 for proteome-wide BPPM at this step.
When 1 ml RIPA is used, the samples should be split into several tubes with < 400 µl of RIPA mixture in each tube.
The fully lysed cells should be translucent. In some cases, adding extra modified RIPA buffer (25% v/v) and another cycle of sonication are required. Keep the samples on ice bath during the sonication to prevent heat-causing denaturing or protein degradation.
This step is necessary for removal of the detergents in the RIPA, which otherwise inhibit subsequent enzymatic labeling reaction.
Transfection efficiency can be confirmed by Western blot at this stage (Gallagher et al., 2008). Once the cells are lysed, the resultant samples should be used immediately. Otherwise, stop at Step 8 in BASIC PROTOCOL 2. For the large scale preparation, we recommend first performing the lysis and Western blot with a small scale of 15 µl RIPA buffer and 1.5 × 105 cells to confirm transfection efficiency.
Upon the transfection of engineered full-length G9a, the resultant cell lysates were treated with Ab-SAM as described below. The proteome-wide substrates of G9a are expected to carry the ‘clickable’ 4-azidobut-2-enyl moiety. These labeled targets were subject to mild copper-free, strain promoted azide-alkyne cycloaddtion (SPAAC) with commercially available fluorescent probe TAMRA-DIBO (Invitrogen). The TAMRA-DIBO-conjugated G9a substrates can then be resolved by SDS-PAGE and visualized with in-gel fluorescence.
The concentration of Ab-SAM and the reaction time described below were customized for the full-length G9a Y1154A mutant in the context of HEK293T cells. These parameters should be optimized for individual PMTs and cell lines of interest. The concentrations of Ab-SAM and the reaction time were evaluated systematically. In general, start with low concentration of cofactors and short reaction time, and gradually increase until desirable signal-to-noise ratios are reached. For certain PMTs with low activities, in-gel fluorescence is not a good choice of BPPM to reveal targets. BASIC PROTOCOL 4 is more suitable for these scenarios.
Using less than 40 µg protein can make subsequent precipitation steps difficult. Steps 1–5 should be carried out with minimal time intervals, otherwise causing the gradual loss of labeling efficiency for unknown reasons.
The purpose of this step is to remove excess cofactor.
TAMRA dye is light sensitive. Keep the samples minimally exposed to light thereafter.
Cover the sample tubes with tin-foil during the heating process to avoid degradation of TAMRA dye. Keep the temperature of the heat block at 70 °C in the denaturing step due to dye degradation at higher temperatures.
Dilute the commercial fluorescent protein ladder at least 10-fold and load the minimal amount for visualization. We also recommend leaving an empty lane between the protein ladder and protein samples to minimize the undesired effect of overflow.
Place the gel box in the dark and cover it with tin-foil to minimize light-mediated dye decomposition.
Use the mode with the excitation at 532 nm, < 580 nm cut-off filter and 30 nm band-pass. Amersham Biosciences Typhoon 9400 was used successfully for this purpose. In general, the background of the gel is high at this stage because of the non-specific labeling from the dye. The background labeling can be significantly reduced through the destaining step.
Steps 11–13 can be repeated multiple times until the images with desired signal-to-noise ratios were obtained.
In case of protein misloading, repeat Step 9 and thereafter after adjusting the protein concentrations
After incubating the cell lysates containing the full-length engineered G9a with Ab-SAM, the substrates of G9a are expected to be labeled enzymatically with the ‘clickable’ 4-azidobut-2-enyl moiety. The samples can be treated with TAMRA-DIBO (Invitrogen) as described in BASIC PROTOCOL 3 for in-gel fluorescence visualization or Click-iT biotin-DIBO (Invitrogen) as described in the current protocol for streptavidin pulldown followed by MS analysis. We consistently observed that MS analysis of BPPM samples reveals more targets than in-gel fluorescence. More importantly, the MS analysis allows the release of protein identities for biological follow-up.
The following conditions have been optimized via in-gel fluorescence visualization as described in BASIC PROTOCOL 3)
This step can also remove excessive cofactor. Steps 1–6 should be carried out with minimal time intervals, otherwise causing the gradual loss of labeling efficiency for unknown reasons.
Solublizing pellets can be facilitated by repetitive pippeting and brief sonication.
This step can remove excess biotin-DIBO probe and thus avoid saturating strepavidin beads in a subsequent step.
Protein samples cannot be fully recovered after the precipitation steps.
The extensive washing steps reduce the amount of proteins non-specifically bound to the beads.
The purpose of this step is to reduce disulfide bonds of cysteines and cap them with the alkylation reagent iodoacetamide to prevent protein aggregation.
We generally use small portion of the samples to run the gel to estimate the amount of the proteins pulled down. Following the confirmation step, we use the remaining sample for proteomic analysis.
We used the following standard protocol to prepare peptide samples for MALDI-MS analysis. The sensitivity of the samples can vary according to the composition of the assay buffer. The presence of detergents in the assay buffer will interfere with subsequent MALDI-MS analysis. If detergent(s) has to be included for the optimal activities of examined enzymes, an alterative protocol should be used to process the samples for MALDI-MS analysis (see ALTERNATIVE PROTOCOL 1).
In our case, desorption/Ionization was obtained by using a 337-nm nitrogen laser with a 3-ns pulse width. Laser power was adjusted to slightly above threshold to obtain good resolution and signal/noise ratios. External standard Calibration Mix 2 (Applied Biosystems) was used for mass calibration. Each measurement was obtained by accumulating three spectra collected at ten different positions with 500 shots per position.
The following is a standard protocol for protein precipitation with a mixture of 4:1:5 (v/v/v) methanol/chloroform/DD water. This precipitation protocol was implemented after click chemistry to remove recessive TAMRA-DIBO probe.
Discard as much of the aqueous layer as possible without disturbing the interface layer, where the precipitated proteins are located. A thin-layer, pink-colored precipitate should be visible between the upper aqueous phase and the lower chloroform phase.
The following is a standard protocol for protein precipitation with methanol. This protocol was used to remove the residual Ab-SAM cofactor after the enzymatic reaction and the excessive Click-iT biotin-DIBO alkyne probe after the click reaction (BASIC PROTOCOL 4). This protocol was developed for large scale protein samples (> 5 mg).
It is important to avoid drying the pellet longer than 20 min, because it is difficult to redissolve fully-dried protein pellet in the subsequent procedures.
The following protocol provides a detailed synthesis, purification, and characterization of Ab-SAM and its precursors. The similar synthesis of Ab-SAM was also described in the previous work (Islam et al., 2011).
In a round-bottom flask, 2 g of (E)-1,4-dibromobut-2-ene (9.34 mM) was placed and dissolved with THF (10 mL). 800 mg of sodium azide (12.2 mM) was dissolved in 1.5 mL water and added to the above mentioned solution of (E)-1,4-dibromobut-2-ene. Reaction was carried out for overnight at ambient temperature (23 °C). Diethyl ether of 100 mL was then added to the reaction and successively washed with 15 mL water and 15 mL saturated NaCl solution. Organic layer was dried with anhydrous Na2SO4 overnight and then evaporated under reduced pressure. Residue was purified by silica gel column chromatography using gradient petroleum ether-dichloromethane (DCM) (1~5%) solvent system to furnish 800 mg of (E)-1-azido-4-bromobut-2-ene (4.6 mM, 49%) and 250 mg of (E)-1,4-diazidobut-2-ene (1.8 mM, 19%) as colorless liquids. 165 mg of starting material (E)-1,4-dibromobut-2-ene (0.77 mM) was also recovered. Rf: 0.25 (5% DCM in hexane) 1H-NMR (500 MHz, CDCl3): δ 6.03-5.97 (m, 1H), 5.85-5.79 (m, 1H), 3.97 (d, 2H, J=7.45 Hz), 3.81 (d, 2H, J=5.85 Hz). 13C-NMR (125 MHz, CDCl3): 130.9, 128.1, 51.6, 30.9.
12 mg of S-adenosyl-L-homocystine (SAH) (0.031 mM) was placed in a capped 4 mL glass vial and dissolved into a freshly prepared 1 mL mixture of formic and acetic acids (1:1) and placed on an ice bath. 538 mg of (E)-1-azido-4-bromobut-2-ene (3.1 mM) and 5.4 mg of AgClO4 (0.031 mM) were added to the acidic solution of SAH. Ice bath was removed after 5 min and the reaction was allowed to warm to ambient temperature (23 °C). Reaction progress was monitored by analytical reversed-phase HPLC (XBridge C18 5 µm 4.6 × 150 mm) at 260 nm eluting with acetonitrile (linear gradient to 10% in 15 min and then to 70% in 5 min) in aqueous trifluoroacetic acid (0.01%) at a flow rate of 1 mL/min. To drive the reaction to completion, the addition of (E)-1- azido-4-bromobut-2-ene (269 mg, 1.55 mM) and AgClO4 (5.4 mg, 0.031 mM) was repeated approximately after 5 h. The reaction was quenched by adding 20 mL of distilled water containing 0.01% TFA (v/v). The aqueous phase was washed with diethyl ether (3 × 15 mL) and then filtered through a Nalgene 0.2 µm syringe filter. Ab-SAM 2 was subsequently purified with preparative reversed-phase HPLC (XBridge Prep C18 5 µm OBD 19 × 150 mm) eluting at a flow rate of 10 mL/min with acetonitrile (linear gradients to 10% in 30 min and then to 70% in 5 min) in aqueous trifluoroacetic acid (0.01%). The diastereomeric mixture of Ab-SAM was collected and concentrated by SpeedVac for 2 h to remove acetonitrile, followed by overnight lyophilization. Ab-SAM 2 was redissolved in water containing 0.01% TFA (v/v) and stored at −80 °C before use. Concentration of Ab-SAM was determined by UV absorption with ε260 = 15,400 L.mol−1.cm−1. The compound as isolated in 40% yield. TR = 10min. 1H NMR (600 MHz, D2O): δ 8.41 (s, 0.5H), 8.39 (s, 1H), 8.38 (s, 0.5H), 5.96−5.91 (m, 0.5H), 5.79− 5.73 (m, 0.5H), 5.67−5.62 (m, 0.5H), 4.76 (q, 1H, J = 5.22 Hz), 4.6 (t, 0.5H, J = 6.54 Hz), 4.58 (t, 0.5H, J = 5.7 Hz), 4.5−4.47 (m, 1H), 4.15 (d, 1H, J = 7.56 Hz), 4.11 (d, 1H, J = 7.56 Hz), 3.88−3.8 (m, 5H), 3.56−3.38 (m, 2H), 2.3−2.28 (m, 2H). 13C NMR (150 MHz, D2O): δ 171.59, 171.49, 163.06, 162.83, 150.15, 148, 144.78, 144.72, 143.5, 143.46, 138.23, 138.06, 119.37, 119.3, 117.92, 117.4, 117.28, 115.35, 90.1, 78.67, 78.41, 73.05 (2C), 72.88, 72.66, 52.11, 52.08, 51.2, 51.15, 42.03, 41.2, 40.81, 35.84, 35.55, 25.3, 25.16. ESI-MS (m/z): 480.1[M]+, 379.09 [5′-(4-azidobut-2-en-1-ynyl)thio-5′-deoxyadenosine + H]+, 250.2 [5′-deoxyadenosine]+. HRMS: calculated for C18H26N9O5S: 480.1778; found: 480.1759.
Although the catalytic domain of G9a and its Y1154A mutant are active in the reaction buffer containing no detergent (BASIC PROTOCOL 1), a small amount of detergent(s) in the reaction buffer can be essential for certain PMTs to be fully active (unpublished results of Luo group). Given that the detergent(s) may interfere with subsequent MALDI-MS analysis, the following protocol was developed to remove the detergent(s).
Deionized distilled water (DD water) was used unless mentioned otherwise.
This buffer was prepared by adding 5 mM freshly-prepared TCEP (the final concentration) and 1 × EDTA-free Roche protease inhibitor cocktail into RIPA buffer (Sigma) prior to use.
This buffer contains 50 mM Tris-HCl, prepared from 1 M stock (pH = 8.0), and 10% (v/v) glycerol. This mixture can be stored at ambient temperature (23 °C) up to 6 months. 2 mM freshly-prepared TCEP (the final concentration) and 1 × EDTA-free Roche protease inhibitor cocktail are added immediately prior to its use.
This buffer containing 40 mM Tris-HCl (pH = 6.8), 69 mM sodium dodecyl sulfate (SDS). 10% (v/v) β-mercaptanol (the final concentration), and 10 % (v/v) glycerol (the final concentration).
This buffer contain 50 mM triethanolamine, prepared from 1 M stock (pH = 7.4), 150 mM NaCl, and 0.5% (w/v) sodium dodecyl sulfate (SDS). This mixture can be stored at ambient temperature (23 °C) up to 12 months. 1 × EDTA-free Roche protease inhibitor cocktail is added immediately prior to its use according to the manufacturer’s instructions.
This buffer was prepared by supplementing GIBCO’s PBS with 0.2 % (w/v) SDS.
Small-molecule chemical reporters coupled with bioorthogonal ligation reactions have recently demonstrated broad application in identifying the substrates of protein posttranslation modifiers (e.g. glycosylation, myristoylation and palmitoylation) (Charron et al., 2009; Prescher and Bertozzi, 2006; Yount et al., 2010). Meanwhile, the active-site engineering for identifying substrates of a specific posttranslation modifier has been successfully documented for various kinases (Bishop et al., 2001; Kapoor and Mitchison, 1999; Lin et al., 2001). The two concepts were merged to develop Bioorthogonal Profiling of Protein Methylation (BPPM) to dissect substrate profiles of designated PMTs. Several previous attempts were made for global profiling of methylation substrates of a specific PMT with radiolabeled SAM as the cofactor and arrayed peptides/proteins as substrate candidates (Levy et al., 2011; Rathert et al., 2008). PMT-knockout proteome was also used as substrate candidates upon profiling PMT substrates (Rathert et al., 2008). However, because of the lack of enrichment strategy after target labeling, these methods are not suitable to detect less abundant methylation targets in biologically-relevant cellular contexts (Luo, 2012). This situation then triggered the development of the SAM analogue cofactors that contain terminal azido/alkynyl moieties (Binda et al., 2011; Dalhoff et al., 2006; Lee et al., 2010; Osborne et al., 2008; Peters et al., 2010; Stecher et al., 2009). These SAM analogues were shown to be active toward certain native PMTs and thus label the PMT substrates with the terminal azido/alkynyl groups. Such labeling provides suitable handles for target enrichment by attaching alkyl/azido-containing biotin probes via the azide-alkyne cycloaddition reaction (CuAAC or click chemistry) (Rostovtsev et al., 2002; Tornoe et al., 2002). Our BPPM technology significantly expands generality and efficiency of the SAM analogues as cofactor surrogates to profile PMT targets in various cellular contexts.
In order to engineer a designated PMT to process SAM analogue cofactors, the PMT construct should be readily expressed and purified exogenously. Native construct of the PMT should show decent activity in vitro toward known peptide substrates, a situation that allows rapid screening by simply replacing the native PMT and SAM cofactor with engineered PMTs and SAM analogues under the same conditions. Optimizing reaction conditions in vitro is critical for establishing the assay parameters to assure the maximal signal when less reactive enzyme-cofactors are used.
The experimental parameters of transient transfection such as cell confluency and the amount of plasmid used can alter significantly between cell types and among PMTs of interest. These parameters need to be optimized for each PMT construct of interest. Overexpression of designated PMTs in certain cell lines could be toxic. In that case, multiple cell lines should be examined to assure optimal transfection efficiency and the balanced toxicity. When working on large scale transfection (using multiple T-75 flasks), we strongly recommend examining a small aliquot of cells to confirm the transfection efficiency with Western blot before proceeding with the protocol(s). The cell lysates should not be stored and need be used immediately. We also recommend using freshly harvested cells for the proposed experiment and often observe the gradual loss of enzymatic activities if the cells are frozen for a long period of time. In general, such cells should not be stored for > 2 months.
It is essential using repetitive gel destaining steps to improve signal-to-noise ratios and thus better visualize the labeled proteins via in-gel fluorescence (BASIC PROTOCOL 3). Residual nonspecifically-bound TAMRA dye in gel lanes leads to high background, which can be removed by multiple washing steps. High quality reagents should be used to prepare the destaining solution to avoid general contaminants. In addition, SDS gels should be handled with great caution only at the edge areas. Physical manipulation of SDS gels leaves fingerprints on the gel, which will become an issue for in-gel fluorescence. Using the recommended amount of proteins is also crucial. Insufficient loading or overloading can cause weak signals or abnormal gel bands, respectively.
For proteome-wide BPPM with MS (BASIC PROTOCOL 4), 5 to 10 mg of Ab-SAM-treated cell lysates are recommended. Performing the protocol with a smaller amount than recommended may limit the detection of low abundant proteins. It is essential to include multiple methanol washing steps after each precipitation step to remove undesired components. After the treatment with Ab-SAM, the washing steps remove the unreacted cofactor, which would otherwise consume the biotin-DIBO probe in the subsequent click reaction. After the click reaction, the methanol washing removes the unreacted biotin-DIBO, which would otherwise compete with biotinylated proteins for the streptavidin beads in the subsequent step of streptavidin pulldown. Fully dissolving precipitated protein pellets is essential and can be achieved by repetitive pipetting with SDS-containing buffers. If this step were an issue, mild sonication could be included. All the steps should be carried out without significant interval unless mentioned otherwise. We noticed that the prolonged intervals from Steps 1–5 (BASIC PROTOCOL 3, 4) cause the gradual loss of labeling efficiency for unknown reasons.
For BASIC PROTOCOL 1, the active enzyme-cofactor pairs are expected to convert the peptide substrate into the corresponding alkylated product (Fig. 3a). The pair of the unmodified and modified peptides will be reflected in MS. To be considered a positive signal, the intensity of the modified peptide needs to account for at least 10 % of the unmodified peptide. For near to 100 % modification, the unmodified peak will disappear completely as shown for G9a Y1154A mutant paired with Ab-SAM (Fig. 3b). The peptide peaks should show the characteristic isotopic distribution, which allows differentiating peptides from noise signals. Magnitude of modification can depend on the amount of cofactor and peptide used. The signal increase correlated with these parameters can be used to further confirm the positive hits.
For the ideal visualization via in-gel fluorescence (BASIC PROTOCOL 3), the control lanes (cells transfected with empty vectors) should show negligible amount of fluorescence labeling in contrast with the sample lanes (cells transfected with the active G9a variant), while the Coomassie staining should be comparable to show that the equal amount of total proteins were loaded (Fig 3c). If both control and sample lanes show weak fluorescence labeling, increasing the amount of loaded proteins in both lanes may result in more pronounced signal. However, it is possible that both control and sample lanes show comparable fluorescence labeling with no distinct band shown in the sample lane. Such situation would indicate either that the engineered PMT constructs are inactive in cellular contexts or that the targets of PMTs in cellular contexts cannot be readily detected via in-gel fluorescence. For either situation, it is worth performing proteome-wide BPPM with MS for further validation, given the higher sensitivity of MS (BASIC PROTOCOL 4). In an ideal situation, the signal lane in contrast to the control should display more labeling bands with Coomassie staining (Fig 3d). In certain situation, such differences cannot be distinguished by Coomassie staining. We therefore recommended performing the proteomic analysis regardless of the results of in-gel fluorescence and Coomassie staining. With the current protocol, we were able to identified > 100 targets, more than what are shown by in-gel fluorescence and Coomassie staining.
BASIC PROTOCOL 1 can be completed within 1 h (depending on the number of samples tested). It takes 3 d to complete the cell culturing and harvest sessions in BASIC PROTOCOL 2. At this stage, the cells can be stored at −80 °C for a few weeks. Preparing resultant cell lysates can be completed in 2 h and must be immediately followed by BASIC PROTOCOL 3 and BASIC PROTOCOL 4. The control experiments using Western blot to confirm the level of the expression can be performed in the same day after harvesting cells. It takes 3.5 h to reach the step of methanol/chloroform/water precipitation and washing, and 1.5 h to resolublize the protein precipitates and run the SDS-PAGE gel in BASIC PROTOCOL 3. The precipitated proteins can be stored at −80 °C for a few days before use. Destaining gel can be performed overnight. It takes 3 d to go through BASIC PROTOCOL 4. Briefly, the enzymatic reaction can be completed within 2 h followed by methanol precipitation. This step can be performed in the same day as the experiment of in-gel visualization in BASIC PROTOCOL 3. Each precipitation step needs to be carried on overnight to assure the high recovery of the proteins. The precipitated proteins can be stored at −80 °C for a few days before use. Subsequent washing step takes around 1.5 h, followed by 1 h click reaction. After redissolving the biotinylated protein pellet, the following steps in BASIC PROTOCOL 4 need to be carried out continuously. Streptavidin pulldown, washing and elution of labeled PMT substrates take 6 h. The resultant proteins can be stored at −80 °C before MS-based proteomic analysis.
The authors thank Prof. Howard Hang for technical supports and financial supports through Mr. William H. Goodwin and Mrs. Alice Goodwin Commonwealth Foundation for Cancer Research, The Experimental Therapeutics Center of Memorial Sloan-Kettering Cancer Center, NIGMS (1R01GM096056), NINDS (R21NS071520), the NIH Director’s New Innovator Award Program (1DP2-OD007335), the V Foundation for Cancer Research (2009 V Foundation Scholar Award, M.L.), March of Dimes Foundation (Basil O’connor Starter Scholar Award, M.L.), Starr Cancer Consortium, and the Alfred W. Bressler Scholars Endowment Fund.