|Home | About | Journals | Submit | Contact Us | Français|
Identifying kinase substrates is an important step in mapping signal transduction pathways, but remains a difficult and time-consuming process. Analog-sensitive kinases (AS-kinases) have been used to selectively tag and identify direct kinase substrates in lysates from whole cells. In this approach a gamma-thiol ATP-analog and AS-kinase are used to selectively thiophosphorylate target proteins. Thiophosphate is used as a chemical handle to purify peptides from a tryptic digest, and target proteins are identified by liquid chromatography and tandem mass spectrometry (LC-MS/MS). Here, we describe an updated strategy for labeling AS-kinase substrates, solid-phase capture of thiophosphorylated peptides, incorporation of stable-isotopic labeling in cell culture (SILAC) for filtering nonspecific background peptides, enrichment of phosphorylated target peptides to identify low-abundance targets, and analysis by LC-MS/MS.
Identification of protein substrates that are phosphorylated by a particular kinase has been time-consuming and prone to experimental artifacts. Analog-sensitive kinases (AS-kinases) have provided a system to label, visualize, and identify kinase substrates (Fig. 1). AS-kinases are engineered with “gatekeeper” mutations in their ATP-binding pocket such that they can accept bulky ATP analogs, generally with a large alkyl group at the N6 position (Fig. 1).1 A wide range of AS-kinases have been generated and extensive validation has shown that the gatekeeper mutation does not affect their substrate specificity.
The Shokat and Clurman groups have each used γ-thiol-substituted ATP analogs to thiophosphorylate substrates of cyclin/cyclin-dependent kinase (CDK) complexes.2, 3 In each case, the labeled substrates have been isolated by solid-phase capture through the thiophosphate group, followed by elution and subsequent identification of captured peptides by liquid chromatography and tandem mass spectrometry (LC-MS/MS). These approaches allow discovery of unknown kinase substrates, but have been challenging to implement. We have developed an updated method to those used by Blethrow et al.2. With this approach, we have identified substrates of the extracellular signal-regulated protein kinase 2 (ERK2) in NIH 3T3-L1 fibroblasts.4 Although a large number of ERK2 substrates were identified, only a small fraction of these substrates had been previously reported, raising the possibility that additional high-throughput screens performed on this and other kinases will reveal signaling networks that are much more interconnected than has been appreciated so far.
This protocol updates earlier procedures in several ways to improve yield, account for nonspecific labeling by endogenous kinases, and reduce nonphosphorylated background so that very low abundance substrates can be identified (Fig. 2). To minimize losses on tube and tip surfaces, the entire process is conducted using capillary columns. The capillary format is also used to perform small-scale phosphopeptide enrichment on immobilized metal affinity chromatography (IMAC) columns.6 This additional phosphopeptide enrichment step substantially reduces nonphosphorylated peptides and is critical for identifying less abundant proteins. Application of Stable Isotope Labeling in Cell Culture (SILAC) provides a quantitative comparison between AS-kinase labeling and the negative control reaction. 5 SILAC also enables a quantitative comparison of substrate abundance between different cells types and conditions. We believe data collected with this methodology will help illuminate how signaling pathways are rewired in different biological contexts.
Note: SILAC reagents are from Thermo Fisher Scientific/Pierce Protein Research Products (Thermo/Pierce).
Note: Other SILAC base medium may be substituted for other cell lines.
Note: This reagent is optional.
Note: This reagent is optional.
Note: High expression and activity of the AS-kinase is critical for effective substrate labeling.
Note: Other stimulating ligands or agonists may be used to activate other AS-kinases.
Note: Sulfur-based reducing agents may not be substituted for TCEP because they are not compatible with thiophosphate capture.
Note: N6-phenyl-2-ethyl gamma-thiol ATP works well for AS-ERK2, another substituent at the N6 position may be better for other kinases.
Note: DTT is used for its nucleophilic thiol groups; other reducing agents may not be substituted.
Note: Other phosphopeptide enrichment strategies, such as titanium dioxide, may be substituted.
Note: An alternative to C18 beads is an HPLC autosampler with in-line reverse-phase cartridge.
Note: An oven set at 80°C may substitute.
Note: A low pressure alternative is a centrifuge column (#89868 or #89896, Pierce/Thermo).
Note: We use an OrbiTrap XL from Thermo Scientific with 20 nL/min flow rate on a home-made tapered tip. Our HPLC gradient is 0% to 70% ACN in 0.2M acetic acid over two hours, and our instrument is set to perform a survey scan with 100,000 resolution followed by up to 10 MS2 experiments using CID fragmentation and analysis in the LTQ ion trap.
Dissolve 1 mg PMA in 1 mL DMSO to make a 1.6 mM stock solution. Add 62.5 μL to 937.5 μL DMSO to make a 100 μM working dilution. Separate the working solution into 100 μL aliquots and store at -20°C for six months.
Dissolve 10 mg of GTP in 191 μL deionized water to prepare a 100 mM stock solution and store 25 μL single-use aliquots at -20°C for up to one year.
Dissolve PNBM at 12 mg/mL in DMSO to prepare a 50 mM stock. Store at -20°C for one year as single use aliquots.
Dilute 25 μL of Recipe 6 by adding 5 μL water five times mixing after each addition (stepwise dilution avoids precipitation). Prepare this solution immediately before use.
Prepare 200 μL per sample of 8M Urea solution by adding 0.48 g/mL of urea to water a volumetric flask or tube. Use this buffer the day it is prepared.
Add 1 mL glacial acetic acid to 1 L deionized water. Store at room temperature.
Add 0.2 mL glacial acetic acid to 20 mL deionized water and 180 mL ACN (90% ACN 0.1% acetic acid). Store in an air-tight glass bottle for up to one year.
Add 0.2 mL glacial acetic acid to 120 mL deionized water and 80 mL ACN (40% ACN in 0.1% acetic acid). Store in an air-tight glass bottle for up to one year.
Add 2 mL isopropanol to 8 mL ACN and store at room temperature in a glass bottle for three months.
Dissolve 1 mg angiotensin peptide in 771 μL 0.1% acetic acid (Recipe 10) to prepare a 1 nmol/μL stock. Dilute the stock 1:1000 in 0.1% acetic acid to prepare a 1 pmol/μL working solution. Store the stock at -20°C for one year; store the working solution at 4°C for one month.
Dissolve 2.4 mg (100 nmol) B-casein in 1 mL Digest Buffer (Recipe 9). Add 20 μg Trypsin, mix well and place at 37°C for six hours to overnight. Dilute the digest 1:100 into 0.1% acetic acid (Recipe 10) to make a 1 pmol/μL solution. Divide into 50 μL single-use aliquots and store at -80°C.
25 mM HEPES, 50% ACN adjusted to pH 7.0
Prepare 100 mL and store in an air-tight glass bottle at room temperature for three months. Discard the solution if precipitate is visible.
Add 5 mM TCEP immediately before use (1:100 dilution from 500 mM stock solution).
Supplement 1 mL of Recipe 16 with 25 μg/mL BSA (1:80 dilution from 2 mg/mL stock) and add 5 mM TCEP immediately before use (1:100 dilution from 500 mM stock solution). Prepare immediately before use and do not store this buffer.
25 mM HEPES, 50% ACN adjusted to pH 8.5
Prepare 50 mL and store in an air-tight container at room temperature for three months.
Add DTT to a final concentration of 5 mM to 0.5 mL of the solution immediately before use.
Add 2.5 mL formic acid to 47.5 mL deionized water (5% formic acid). Store in an air-tight glass bottle for three months.
Dissolve 10 mg Oxone to 1 mL deionized water to prepare a 10 mg/mL stock solution, and then add 200 μL of the stock solution to 800 μL deionized water to yield a 2 mg/mL working solution. Prepare both solutions immediately before use.
Dissolve 292 mg of EDTA in 10 mL deionized water (100 mM) and adjust the pH to 8.9. Store in an air-tight glass bottle for three months.
Dissolve 1 g iron (III) chloride in 61 mL deionized water [100 mM iron (III) chloride].
Store in an air-tight glass bottle for up to one month. Prepare a fresh solution if any precipitate is visible.
Add 0.5 mL glacial acetic acid and 12.5 mL ACN to 47 mL deionized water (1% acetic acid 25% ACN). Store in an air-tight glass bottle for three months.
Prepare 50 mL of 250 mM sodium phosphate with pH adjusted to 9.0. Store in an air-tight glass bottle for three months.
Add 12 mL glacial acetic acid to 988 mL HPLC-grade water (0.2 M acetic acid) and store in a glass bottle for six months.
Add 700 mL ACN and 12 mL glacial acetic acid to 288 mL HPLC-grade water (0.2 M acetic acid, 70% ACN) and store in a glass bottle for six months.
We describe the conditions for SILAC labeling 3T3-L1 cells expressing wild-type and AS-ERK2. Growth medium and conditions for serum-starvation and stimulation can be adjusted to analyze other kinases. The kinase reaction should be performed with 4-6 mg of protein in approximately 600 μL of Lysis/Kinase Reaction Buffer (Recipe 4). This concentration allows in vitro labeling to proceed efficiently, and matches the capacity of the Sep-Pak cartridge. Western blotting can be used to determine the extent of thiophosphate labeling.
Note: Appropriate SILAC base medium can be selected for the cell-type of interest. We generally grow one 15-cm plate to 80% confluence for the labeling reaction.
Note: AS-ERK2-expressing 3T3-L1 cells only tolerate 3 hour serum-starvation. Serum starvation attenuates the activity of the kinase being studied prior to the labeling reaction.
Note: This step rapidly actives the ERK2 kinase while minimizing the time available for substrate phosphorylation. Use an appropriate stimulation reagent for the cell type and kinase of interest.
Note: The lysis buffer is very mild and leaves a considerable amount of cell debris. Debris may contain kinase substrates and should not be removed from the reaction.
Note: This step is optional and is used for troubleshooting purposes.
Note: If desired for troubleshooting purposes, 13.5 μL aliquots can be taken every 10-15 minutes for analysis by Western blotting. Stop the reaction by adding EDTA to 50 mM and placing the samples ice.
Note: This step is optional. Western blotting is used to estimate the extent of phosphorylation before and after thiophosphate labeling reactions. Additional samples taken during step 11 can be used to optimize the time for the labeling reaction to minimize background labeling.
Note: This step is optional, but it may be helpful for troubleshooting the thiophosphate labeling.
The proteins in the lysates are precipitated to remove detergents, then denatured with urea, followed by digestion with trypsin. We describe precipitation with chloroform and methanol, but other protein precipitation methods, such as precipitation by ammonium sulfate, can also be used. Precipitated proteins should be physically disrupted by pipetting, vortexing, or sonication, but they will not dissolve completely until after digestion with trypsin.
Note: Chloroform is hazardous, use in a chemical fume hood.
Note: The pellet will not disperse completely until it is digested with trypsin.
Digested proteins must be desalted prior to thiophosphate capture. We use C18 Sep-Pak cartridges. Sep-Pak Plus cartridges can accommodate 4 mg of digested peptide; Sep-Pak Light cartridges should be used for samples with less than 1 mg of digested peptide. Divide the digest as necessary so that no more than 4 mg of peptide is applied to each Sep-Pak cartridge.
Note: Acidified samples may be stored indefinitely at -80°C.
Note: We recommend using a ring stand to suspend the syringe over a waste container during this process.
Note: Samples may be stored indefinitely at -80°C.
Note: Lyophilized samples may be stored indefinitely at -80°C.
Substrate peptides eluting from the SulfoLink beads are captured with a capillary column packed with POROS R2 bead (Fig. 3). After washing to remove Oxone from the R2 column, peptides are transferred to a capillary column containing IMAC beads charged with iron (III) to further enrich phosphorylated peptides. Finally, peptides are transferred from the IMAC column to a capillary column containing C18 beads. The C18 column is then placed in-line for identification of substrate peptides by LC-MS/MS. We use the POROS R2 column for the initial reverse-phase capture and rinsing because it supports a flow rate of 4 μL/min at pressure below 200 PSI, whereas the C18 column provides better chromatographic separation during LC-MS/MS. One of each capillary column should be prepared before proceeding with the protocol. Notes in the protocol indicate alternative approaches if capillary columns are not available.
Note: ACN flow should be visible but it should not spray. Briefly apply additional heat to the frit if it does not flow freely with pressure below 100 PSI.
Note: After this step, the process for finalizing the columns diverges, with the R2 and C18 columns handled similarly and the IMAC column differently.
We block nonspecific binding sites on the beads with angiotensin, which is an XX amino-acid peptide. This also reduces sample loss. Other standard peptides may be used as an alternative to angiotensin.
Note: Blocked and finalized columns may be stored at XX for XX.
We condition IMAC capillary columns prior to use by loading and eluting 5 picomoles of digested B-casein. This blocks nonspecific binding sites and improves recovery and specificity of the phosphopeptide enrichment. Immediately prior to use a conditioned IMAC capillary column must be loaded with iron (III) chloride. We recommend beginning this process during step 16 of the next section.
Note: ACN flow should be visible but it should not spray. Briefly apply additional heat to the frit if it does not flow freely with pressure below 100 PSI.
Note: The conditioned IMAC column may be stored for three months at room temperature.
We use SulfoLink beads to capture thiol-containing peptides, including both thiophosphate-labeled substrate peptides and cysteine-containing peptides. These beads are light sensitive until after the quenching step and must be protected from light until that step.
Note: The buffer pH will drop when sample peptides are brought into solution. The buffer pH should not be readjusted at this step.
Note: SulfoLink beads degrade once they have been exposed to air and should remain sealed until immediately before use (discard the remainder, do not reseal).
Note: The stronger centrifugation is necessary because the Formic Acid Wash is more viscous than the other binding and washing buffers.
We use a helium packing device to transfer SulfoLink beads loaded with labeled substrate peptides to a capillary column for additional washing and elution to a POROS R2 column (Fig. 3). Thiophosphorylated peptides are converted to normal phosphate during the elution step, making them suitable for phosphopeptide enrichment. If capillary columns are not available the beads may instead be transferred to a 2 mL centrifuge column and washed by gravity-flow. Details about how this changes the procedures are indicated in notes associated with specific steps where necessary.
Note: Alternatively,bake the frit for 15 minutes at 80° C.
Note: This step requires extremely low pressure to avoid spraying ACN from the column.
Note: One end of the tubing will be drilled to match the gauge of the SulfoLink capillary, and the other will be drilled to match the smaller POROS R2 column.
Note: From this point all buffers flowing over the SulfoLink beads will also pass through the POROS R2 column. This allows peptides to be captured as they are eluted while minimizing losses from sample handling or adsorption of peptides onto surfaces. If capillaries are not being used, peptides may be eluted by incubating SulfoLink beads in one bead volume of Oxidizing Elution Buffer (Recipe 20) for five minutes at room temperature with rotation. Eluted peptides should be further enriched by desalting on a reverse-phase surface followed by phosphopeptide enrichment by IMAC (desalting is necessary because Oxone is incompatible with phosphopeptide enrichment), or else eluted peptides may be desalted and analyzed directly by LC-MS/MS.
Note: It is important to flush the entire volume of the column so that all peptides are collected on the R2 column. The duration of this step can be adjusted as necessary.
Note: This step transfers peptides from the R2 to the IMAC column and a slow flow rate is critical so that phosphopeptides have time to bind to the IMAC column.
Note: Alternatively, peptides may be eluted from the IMAC column without a C18 column attached and loaded manually or with an autosampler onto an HPLC-MS/MS system. There is very little peptide at this stage so take care to avoid adsorption on surfaces or loss during transfers.
Note: The IMAC column may be stored and reused for up the three months.
Note: The C18 column is ready to be placed in-line for peptide identification by LC-MS/MS.
Note: The LC-MS/MS protocol will depend on the MS equipment and configuration. We recommend consulting an experienced mass spectrometry facility about how to analyze the samples. Peptides phosphorylated on serine or threonine often have a dominant neutral loss of 98 Dalton. Fragmenting the neutral loss ion with MS3 or multistage activation may improve peptide identification.
Note: Because of the oxidizing elution conditions, methionine residues are always oxidized to sulfoxide or sulfone. Follow-up experiments on a previously unknown substrate are time-consuming and difficult; therefore, we recommend checking automatic assignments by manually assigning every peak to a predicted fragment ion.
Note: For example, use the AS-kinase sample versus the wild-type sample at three standard deviations above the mean for nonphosphorylated peptides.
First, verify that the activated AS-kinase is present in the sample by either Western blotting for phosphorylation on the activation loop (if applicable to the kinase) or adding a recombinant substrate protein and checking for thiophosphorylation by in vitro kinase reaction. If the kinase is present and active, determine whether the AS-kinase accepts thiophosphate by performing kinase reactions with purified AS-kinase and substrate protein in the presence of gamma-thiol ATP. Check for activity by Western blotting for thiophosphate. Some kinases do not accept gamma-thiol ATP analogs. If the AS-kinase accepts thiophosphate, then determine whether the AS-kinase accepts the particular ATP analog used in the experiment. A number of N6-substituted gamma-thiol analogs are available and it may be necessary to optimize the reaction by empirically testing various analogs with the AS-kinase of interest.
It is possible that a weak signal is due to extensive phosphorylation prior to cell lysis, which would leave few sites available for the AS kinase reaction and thiophosphate labeling. Western blotting or MS analysis of the samples before and after the kinase reaction should provide an indication of the extent of phosphorylation or lysate. If the AS-kinase is active prior to lysis, then many of its substrates will be phosphorylated and not available for thiophosphate labeling.
Excess, nonspecific labeling is the most common cause of background signal in the negative control reaction. One was to minimize this problem is to lower the concentration of ATP analog to reduce nonspecific utilization by other kinases. Conducting a time-course experiment by taking aliquots from the labeling reactions every 10-15 minutes can aid in determining the optimal duration of the labeling reaction.
Sometimes the aliquots taken after labeling for Western blotting may show a strong signal, yet the MS data does not have labeled peptides. Loss of signal in the MS samples can result from inefficient thiophosphate capture. The pH of the peptide sample in Thiophosphate Binding Buffer with BSA (Recipe 17) should be between 5 and 5.5 after addition of peptide in order for the reaction with the SulfoLink beads to proceed efficiently. If the pH is too high then cysteine competes for binding to the beads. The pH of the peptide solution may be adjusted with dilute HCl if necessary (we recommend adjusting the pH of an aliquot of binding buffer to 5.0 and then adding the necessary volume). Inefficient capture due to degradation of the SulfoLink beads can also cause a loss of the signal in the MS data. The SulfoLink beads should not be unsealed until immediately before use.
Loss of signal in the MS samples can result from using TCEP that is too old or that was added to the buffers too early. TCEP stock solution should be replaced every three months. Loss of the MS signal can also result from inefficient phosphopeptide enrichment. To verify this process, digest a known phosphoprotein, such as B-casein, and perform phosphopeptide enrichment followed by LC-MS/MS to ensure that phosphopeptides are efficiently recovered.
A thiophosphate positive control can be generated using gamma-thiol ATP (no alkylation at N6) with a recombinant kinase and its substrate, such as Jun N-terminal kinase 1 (JNK1) and c-Jun. Spike 10 picomole of thiophosphorylated c-Jun into the AS-kinase substrate labeling reaction before the precipitation step. Use this to verify that the known substrate peptides are recovered and to determine an approximate limit of detection the protocol.
This protocol is derived from the Shokat lab protocol published in Blethrow et al.2 An alternative capture chemistry was described by the Clurman lab in Chi et al. in which thiophosphorylated peptides are bound to a bead surface by formation of a disulfide linkage.3 Treatment with base selectively hydrolyzes the phosphorous-sulfur bond of the thiophosphate, releasing the bound peptides with normal phosphate in place of the labeled group. Cysteine-containing peptides are bound through a thioether bond that is not affected by base. In our hands, the Clurman protocol suffered from substantial binding of nonphosphorylated peptides, but we believe that the protocol could be optimized to work effectively. As an alternative to capturing labeled peptides, thiophosphorylated proteins may be alkylated with PNBM and immunoprecipitated using the thiophosphate ester antibody, followed by SDS-PAGE and identification by LC-MS/MS.1, 9
Even with efficient substrate labeling, capture, and identification; these approaches still have substantial limitations. The LC-MS/MS approach used here is only suitable for phosphorylation sites that fall within tryptic peptides suitable for the analysis. Many kinases include arginine or lysine in their recognition motif and may produce peptides too short for identification. Alternative proteases may be used as appropriate. Peptides containing cysteine are also likely to be captured on SulfoLink beads through the cysteine thiol instead of through thiophosphate. We have not observed any cysteine-containing peptides that we believe to be bona fide substrate sites.
The protocol remains limited by the requirement to use in vitro kinase reactions in cell lysate instead of performing labeling in intact cells. We have experimented extensively with permeabilization strategies to get ATP analogs into the cell. Although Banki et al. have reported in-cell labeling using permeabilized cells9, we have achieved very limited labeling in cells, sufficient for identification of the most abundant sites, but not at an amount suitable for low-level substrate identification. An alternative approach would to be to conduct labeling reactions in isolated organelles (note that the low-detergent lysis buffer in this protocol may leave protein complexes and some organelles intact).
There is also a biological limitation in that the kinase substrates must not be phosphorylated when the cells are lysed. This is especially challenging in cases with high endogenous kinase activity (for example, kinase substrates downstream of constitutively activating mutations). These systems require extensive optimization to reversibly inhibit kinase activity long enough for substrates to be dephosphorylated, followed by rapid reactivation of the kinase prior to lysis.
Despite all limitations and caveats, unbiased identification of kinase substrates can reveal a huge number of novel targets. Application of this approach to a variety of kinases and biological systems has the potential to reveal previously unknown components of signaling networks.