|Home | About | Journals | Submit | Contact Us | Français|
The majority of eukaryotic proteins are phosphorylated in vivo and phosphorylation may be the most common regulatory post-translational modification. Many proteins are phosphorylated at numerous sites, often by multiple kinases, which may have different functional consequences. Understanding biological functions of phosphorylation events requires methods to detect and quantify individual sites within a substrate. Here we outline a general strategy that addresses this need and relies on the high sensitivity and specificity of selected reaction monitoring (SRM) mass spectrometry, making it potentially useful for studying in vivo phosphorylation without the need to isolate target proteins. Our approach uses label-free quantification for simplicity and general applicability, although it is equally compatible with stable isotope quantification methods. We demonstrate that label-free SRM-based quantification is comparable to conventional assays for measuring the kinetics of phosphatase and kinase reactions in vitro. We also demonstrate the capability of this method to simultaneously measure relative rates of phosphorylation and dephosphorylation of substrate mixtures, including individual sites on intact protein substrates in the context of a whole cell extract. This strategy should be particularly useful for characterizing the physiological substrate specificity of kinases and phosphatases, and can be applied to studies of other protein modifications as well.
The function of most proteins is regulated in some way by enzymatic post-translational modifications (PTMs). A large number of covalent PTMs exist in nature and can affect many aspects of protein function including enzymatic activity, binding partners, stability, structure, and cellular localization. Most PTMs are reversible, and therefore allow dynamic control of protein function. Reversible phosphorylation of serine, threonine, and tyrosine, controlled by the opposing activities of kinases and phosphatases, is the most abundant and heavily studied regulatory PTM in eukaryotic cells. At least 75% of human proteins are phosphorylated in vivo  and many proteins are phosphorylated on multiple sites that can be subject to differential regulation , often by more than one kinase and/or phosphatase. As an example, more than 100 in vivo phosphorylation sites have been mapped on the human anaphase-promoting complex (APC), a large E3 ubiquitin ligase complex that regulates progression through mitosis [1,3-4]. Evidence suggests that the APC is regulated, both positively and negatively, by at least four different kinases [5-12]. This complexity makes functional studies of protein phosphorylation challenging. There is a need for improved methods for detection and quantification of individual PTM sites on multiply modified proteins.
Very few analytical methods employed to study phosphorylation, or PTMs in general, are effective at quantitatively detecting multiple sites independently on the same protein. Mass spectrometry (MS) is one exception. Reduction of proteins to peptides is an effective way to separate the PTMs on a single protein into distinct molecules for mass spectral analysis. Tandem MS can provide the sequence of individual peptides and the exact location of any PTMs. Although the complexity of biological systems poses a constant challenge to the detection and quantification of target molecules, the development of affinity enrichment strategies, multi-dimensional chromatographic separations, improved instrument designs, and various other methodological advances are steadily allowing more comprehensive study of the proteome. Selected reaction monitoring (SRM) MS, which relies on the filtering capabilities of a triple quadrupole mass spectrometer, has become a popular method for direct detection of specific target proteins within complex biological samples. Because of its sensitivity and specificity, SRM is also a useful approach for the quantitative study of PTM regulation in living systems.
To fully understand regulation and function of PTMs it is necessary to measure their dynamic changes under different conditions and in response to physiological and environmental signals. A variety of methods exist for the absolute and relative quantification of proteins and PTMs from peptide-based mass spectral data [13-14]. Most quantitative MS approaches employ some form of stable isotope dilution, which has the advantage of allowing simultaneous analysis and comparison of multiple samples or a single sample with an internal standard [15-17]. However, use of stable isotopes is not essential for quantification and a number of “label-free” approaches have also been developed . Label-free quantification requires comparison of separate MS experiments, which necessitates exquisite care in sample preparation and the appropriate controls or standards to account for run to run signal variations. Although stable isotope-based methods are generally thought to be more accurate than label-free approaches , there are important advantages to label-free methods such as simplicity and general applicability. With label-free quantification there are no specialized reagents required, no extra chemical reaction and sample processing steps, and the methods are suitable for any experimental system. For these reasons, label-free quantitative MS methods have important advantages for the study of PTMs and are becoming more common.
A variety of strategies have been developed for quantitative MS-based analysis of phosphorylation and other PTMs (for example [18-24]). The majority of these methods rely on stable isotope labeling or addition of specific standards for quantification. Some are limited to studies of synthetic peptide substrates or highly purified protein preparations or require phosphopeptide enrichment steps. We set out to establish a strategy that will be generally useful for the quantitative measurement of changes in phosphorylation on intact proteins. Our approach uses label-free quantification of phosphorylation stoichiometry [21,25-26] for simplicity, compatibility with all experimental systems, and practicality for studies of large numbers of phosphorylation sites. We used SRM MS of peptides generated by enzymatic digest because the sensitivity, specificity, and dynamic range of SRM  make it suitable for analysis of highly complex biological samples, and it is also capable of simultaneous monitoring of up to hundreds of target molecules in a single sample run. Related SRM MS methods were developed and used recently. In one case, changes in phosphorylation of the Pho4 transcription factor in budding yeast were measured , but without a rigorous quantification strategy. In another case, autophosphorylation sites on the Lyn tyrosine kinase in mouse xenograft tumors were monitored . Our strategy is equally useful for studying phosphatases and kinases, and can be applied in vitro as an enzymatic assay, or in vivo to quantify changes in many phosphorylation sites in response to experimental variables. Among other potential applications, we predict it will be very useful for the exploration of kinase and phosphatase specificity. The general method is also suitable for studying other PTMs.
In this report we describe proof of principle experiments that validate this strategy. We demonstrate that SRM MS combined with label-free quantification is comparable to conventional assays for measuring reaction rates and steady-state kinetic parameters of phosphatases and kinases. We also demonstrate the utility of the method for simultaneously monitoring multiple phosphorylation sites on intact protein substrates, including in a whole cell extract, and the potential to reveal different responses of individual sites to experimental variables. Possible experimental designs and practical aspects of applying this method to large-scale studies in complex biological mixtures and in vivo are discussed.
Peptides were synthesized using CLEAR-Amide resin (Peptides International) at 50 μmol scale by solid-phase FMOC chemistry on a Prelude peptide synthesizer (Protein Technologies, Inc.) essentially as described previously . The only modification was that coupling times for FMOC-phosphoserine and -phosphothreonine (Anaspec, Inc.) were increased to 3 hours. Unphosphorylated FMOC-protected amino acids were from Peptides International. Synthesis products were resuspended in 5% acetonitrile/0.1% TFA, fractionated by HPLC using a preparative C18 column (Agilent Technologies) and evaluated by MALDI-TOF mass spectrometry (Voyager 4800, Applied Biosystems). Fractions exhibiting the highest product purity were pooled, lyophilized, and resuspended in 50 mM Tris pH 8.0 at approximately 5 mM based on weight. Precise concentrations of phosphopeptides were determined by measuring the inorganic phosphate released after ashing in a magnesium nitrate solution and reaction with an ammonium molybdate/malachite green mixture as described . Unmodified peptide concentrations were determined by amino acid analysis at the Purdue Proteomics Facility.
The Saccharomyces cerevisiae FIN1 gene was cloned into pGEX-6P-1 (GE Healthcare) to create an in-frame fusion with the vector GST sequence. The N-terminal fusion protein GST-Fin1 was overexpressed in a 3 L E. coli culture by 0.5 mM IPTG induction at 37° C for roughly 4 hours. Unless otherwise stated, all purification steps were performed at 4 °C. Washed cells were suspended in 5 pellet volumes of 50 mM HEPES pH 8.0, 1% Triton X-100, 250 mM NaCl, 1 mM EDTA, 10% glycerol, 1 mM phenylmethlysulfonyl fluoride (PMSF), 100 μM leupeptin, and 1 μM pepstatin, treated with 1 mg/ml lysozyme on ice for 30 min and lysed by sonication on ice. The extract was clarified by centrifugation at 35,000 x g for 30 min and then incubated with 500μL of GST-Bind Resin (EMD Biosciences) for 30 min. Unbound proteins were then removed by extensive washing with 50 mM HEPES pH 8.0, 0.1% Triton X-100, 250 mM NaCl, 1 mM EDTA, and 10% glycerol. GST-Fin1 was then eluted with the same buffer supplemented with 10 mM reduced glutathione and dialyzed against 10 mM HEPES pH 7.5, 10 mM MgCl2, 50 mM NaCl, 0.5 mM DTT and 10% glycerol overnight. 6His-Cdc14 was purified as previously described . Following purification, pooled Cdc14 fractions were exchanged into 50 mM Tris-HCl pH 7.5, 300 mM NaCl, 2 mM EDTA and 0.1% β-mercaptoethanol using a G-25 desalting column.
Saccharomyces cerevisiae Cdk1 (Clb2-Cdc28) was purified from yeast strain BY4741 sic1Δ overexpressing a Clb2-protein A (Clb2-PrA) fusion protein from BG1805-CLB2 (Open Biosystems). Clb2-PrA expression was induced for 4 hours at 30 °C by addition of 2% galactose to the YP-raffinose medium when cells had reached mid log phase. Washed cells were lysed in 5 pellet volumes of 50 mM Tris-HCl pH 8.0, 250 mM NaCl, 10% glycerol, 0.1% Triton X-100, 20 mM NaF, 1 mM PMSF, 100 μM leupeptin and 1 μM pepstatin by disruption with glass beads in a Bead Beater (Biospec Products). Extracts were clarified by centrifugation at 35,000 x g for 30 min and incubated with 150 μl IgG-agarose beads (Sigma) for 2 hours at 4°C. The IgG resin was collected by gentle centrifugation and washed 4x with 50 mM Tris-HCl pH 8.0, 250 mM NaCl, 10% glycerol, 0.1% Triton X-100, and 20 mM NaF. For peptide kinase assays the Clb2-PrA was eluted by cleavage with PreScission protease (GE Healthcare) overnight at 4 °C as recommended by the supplier and the GST-tagged protease subsequently removed by incubating the eluted material with 10 μl GST-Bind resin (EMD Biosciences) for 15 min followed by centrifugation to pellet the resin. The supernatant containing purified Cdk1 was directly stored at −20 °C. For phosphorylation of recombinant GST-Fin1, the IgG agarose beads were washed with kinase buffer (see below), suspended in kinase buffer supplemented with 30% glycerol and stored directly at −20 °C without elution.
S. cerevisiae cells were grown to mid log phase in standard rich medium, harvested, washed, and lysed in the same buffer described above for Cdk1 purification by glass bead disruption. The clarified extract (50 μg) was subjected to SDS-PAGE and Coomassie blue staining. The entire lane was excised from the gel in small pieces and proteins digested with trypsin (see below). Extracted peptides were pooled to create the digested lysate sample for evaluation of GST-Fin1 dephosphorylation in a complex mixture.
Kinase assays (20 μl) contained 2 μL of affinity-purified Cdk1, varying concentration of synthetic peptide substrate, and 1 mM ATP in kinase buffer (10 mM HEPES pH7.5, 10 mM MgCl2, 50 mM NaCl, 10% glycerol, and 0.5 mM dithiothreitol). Duplicate reactions were performed, one of which contained [γ-32P]-ATP (MP Biomedicals). Reactions were initiated by addition of ATP and incubated at 30 °C for 20 min. [γ-32P]-ATP reactions were terminated with 1% trifluoroacetic acid. Cold reactions were terminated by diluting the sample 5-fold with 3% acetonitrile in 0.1% formic acid. [γ-32P]-peptide products were purified from free ATP using OMIX-C18 micropipet tips (Varian, Inc., loading capacity 8 μg) according to the manufacturer’s protocol. After washing, the portion of the tip containing the C18 resin was removed and the resin subjected to liquid scintillation counting. Product formation in the cold ATP reactions was monitored by SRM MS as described below.
Recombinant GST-Fin1 was phosphorylated in vitro by incubating 350 μl of stock protein with ~50 μl Cdk1 resin for 30 min at 30 °C. The reaction was stopped by pelleting the resin, removing the supernatant, and storing in aliquots at −80 °C.
Synthetic phosphopeptide dephosphorylation reactions contained 50 nM Cdc14, varying concentrations of phosphopeptide substrate, 50 mM Tris-HCl pH 7.5, 150 mM NaCl, and 1 mM EDTA. Reactions were incubated at 30 °C and stopped at the desired times by addition of an equal volume of 1.2 M HCl. Reaction products were divided in half. One half was diluted to 1 pmol/μL with 3% acetonitrile in 0.1% formic acid and analyzed by SRM MS as described below. The other half was analyzed spectrophotometrically by measuring A640 after mixing with a solution of ammonium molybdate and malachite green as described . Total inorganic phosphate produced was determined from a Na2PO4 standard curve.
Phosphorylated recombinant GST-Fin1 was dephosphorylated by addition of 200 nM Cdc14. The reaction was incubated at 30°C and aliquots removed at 2, 5, 20, and 60 min and stopped by boiling in SDS-PAGE loading dye. Reaction products were subjected to SDS-PAGE with Coomassie blue staining and the GST-Fin1 band excised, destained with 50% acetonitrile/50 mM ammonium bicarbonate, dried with pure acetonitrile, and digested in-gel with either 10 μg/ml Lys-C or 20 μg/ml trypsin in fresh 50 mM ammonium bicarbonate overnight at 37 °C. Peptides were extracted twice with acetonitrile, dried, and resuspended in 3% acetonitrile/0.1% formic acid immediately prior to MS analysis.
SRM experiments were conducted on an Agilent 6410 triple quadrupole mass spectrometer equipped with an HPLC-Chip cube interface and a 150 mm C18 ProtID-Chip connected to an Agilent 1200 series nanoflow HPLC system. SRM method parameters for all peptides monitored in this study, including parent ion charge state, collision energy, and fragment ion m/z values were manually determined and are listed in Table 1. Collision energies (CE) were first estimated using the formulas CE = 0.044 × m/z + 5.5 (charge state = +2) and CE = 0.051 × m/z + 0.5 (charge state = +3) , and then further refined. At least two transitions were monitored for each peptide and were validated with product ion scans (not shown). Synthetic peptides were eluted from the C18 column and directly injected into the 6410 with a 6 minute gradient of increasing acetonitrile from 3 to 35% in 0.1% formic acid at a flow rate of 400 nL/min. Samples containing proteolytic peptides from recombinant GST-Fin1 were eluted from the C18 column with a similar gradient extended to 20 minutes. Resolution for quadrupoles 1 and 3 was set to wide for synthetic peptide analyses and to unit and wide, respectively, for recombinant protein analyses. The typical source voltage was 1900-2000 V and other instrument settings were ΔEMV = 300, gas temp = 300 °C, and gas flow = 4 L/min.
All raw SRM signals were integrated using Agilent’s MassHunter software. Standard peptides were included in all reactions to correct for run to run signal variation. For analyses of recombinant GST-Fin1, additional proteolytic peptides with strong signal intensity, no Cdk consensus phosphorylation sequences, and no detectable missed cleavages were used as internal standards. The average percent deviation of all standard peptide signals in each sample being compared was determined as described  and used to adjust the raw SRM signals. Moles of product formed in the synthetic peptide phosphatase assays were determined from phosphorylation stoichiometry calculations obtained by comparison of unmodified and phosphorylated peptide SRM signals  and the known starting substrate concentration. The kinetic parameters KM and kcat were determined by measuring product formation as a function of substrate concentration and fitting the Michaelis-Menten equation to the data using GraphPad Prism software.
Our goal was to establish an approach to measure the activities of kinases and phosphatases on their natural physiological substrates in a site-specific manner that could be applied to in vitro and in vivo studies with any experimental system. To facilitate use of the approach in complex mixtures such as cell extracts or to monitor activities in vivo we used SRM MS on a triple quadrupole mass spectrometer. We also chose a simple label-free method for determining PTM stoichiometry for quantification [21,25-26]. Although SRM has been used previously to measure enzymatic activities, its use in measuring kinase and phosphatase activities, or changes in phosphorylation stoichiometry under different experimental conditions, has been very limited. To our knowledge measurement of phosphatase and kinase activities by SRM MS, particularly with label-free quantification, has not been validated by comparison to conventional assays. We therefore set out initially to do so using a defined system with synthetic peptide substrates and purified enzyme preparations.
Consistent with other SRM studies using triple quadrupole instruments, product detection in our phosphatase reactions was linear over a 10,000-fold concentration range with sub-femtomole sensitivity (Figure 1A). The dynamic range was limited only by the capacity of the HPLC column used and therefore may be even greater. The sensitivity was typical for our synthetic peptide substrates but varies substantially with sequence. Precision of SRM measurements was tested by repeat analysis of single samples. Run to run variation was consistently less than 4 percent (Figure 1B). Satisfied with these basic properties of the SRM analysis we next compared initial dephosphorylation rates of several phosphopeptide substrates of the yeast Cdc14 phosphatase measured using a malachite green dye-based colorimetric assay and our SRM method to determine if SRM MS is capable of accurate kinetic measurements. Individual completed reactions were split and products analyzed by the two methods (using a spectrophotometer and mass spectrometer) in parallel.
Figures 1C-E show overlay plots of dephosphorylation versus time for three phosphopeptide substrates measured using the two independent methods. The plots and calculated rates of peptide dephosphorylation were very similar. In our experiments we used our previously described method to determine the percentage of peptide dephosphorylated at each time point from the substrate and product SRM signals  and converted this to moles of product formed using the known initial substrate concentration. Other approaches are also possible when absolute quantification of product is required, including use of stable isotope labeled product as an internal standard.
We also compared the ability of the conventional and SRM-based phosphatase assays to measure reaction velocity as a function of substrate concentration for determination of the steady-state kinetic parameters KM and kcat (Figure 2A). Consistent with the rate measurements, overlays of velocity versus substrate concentration plots generated from the independent methods were in close agreement and calculated KM and kcat values obtained by nonlinear regression analyses were very similar (Table 2). To demonstrate that the approach is equally useful for measuring kinase activity, we measured KM values for budding yeast Cdk1-catalyzed phosphorylation of an unmodified peptide containing a consensus Cdk phosphorylation sequence using conventional 32P incorporation compared to SRM MS (Figure 2B). Again, the plots and calculated KM values were very similar. These results demonstrate that SRM MS can be used for the accurate measurement of phosphatase and kinase activities in vitro.
Since the power of SRM MS lies in its ability to quantitatively measure many specific target molecules within a complex mixture, we next demonstrated the simultaneous measurement of a mixture of phosphopeptide substrates dephosphorylated by Cdc14 (Figure 3). The relative rates of dephosphorylation (with each peptide’s initial SRM signal normalized to 100%) of seven of our synthetic phosphopeptides were measured. Note that most conventional assays for phosphatase and kinase activity are not useful for analysis of substrate mixtures and provide only global data with multiply phosphorylated substrates because they do not distinguish between individual sites. Although the specificity of SRM is not necessarily required for measurement of synthetic peptide mixtures of this complexity (standard MS analyses would suffice, particularly with high resolution instruments), this approach would allow rapid screening of large peptide mixtures for characterization of enzyme specificity or identification of potential enzyme substrates. Similarly, the approach can be applied to analysis of many phosphorylation sites on intact protein substrates in a cell extract following post-reaction enzymatic digest without the need for affinity purification or fractionation steps. In these cases, the specificity and sensitivity of SRM MS is essential.
As an example of the utility of this approach, we simultaneously monitored multiple phosphorylation sites on an intact protein substrate of Cdc14, the yeast spindle stabilizing protein Fin1 . Recombinant GST-Fin1 was phosphorylated in vitro by purified Cdk1 as described in Materials and Methods. Cdk phosphorylation sites are the preferred substrates of Cdc14 phosphatases [34-35] and previous studies have demonstrated that Fin1 is regulated by both Cdk and Cdc14 . Intact phosphorylated GST-Fin1 was reacted with Cdc14 and prepared for MS analysis by protease digestion. The resulting peptide mixture was analyzed by SRM MS. We monitored both the unmodified and phosphorylated forms of 4 peptides containing consensus Cdk sites (S/T-P) that are known in vivo phosphorylation sites of biological importance [33,36-38], as well as several additional peptides lacking Cdk1 sites that were used for normalization. Estimated initial phosphorylation stoichiometry was similar for each site and plotting relative phosphorylation of the individual Fin1 Cdk1 phosphorylation sites over time after Cdc14 addition revealed an unexpected difference in apparent dephosphorylation rates (Figure 4A). These results suggest that not all Fin1 Cdk phosphorylation sites are equivalent substrates for Cdc14 and that additional specificity determinants may exist. This type of observation can only be made using an assay capable of quantifying phosphorylation at individual sites and highlights the potential of our method to study kinase and phosphatase specificity using intact multiply-phosphorylated and physiologically relevant protein substrates, including large protein complexes, and in the context of complex mixtures such as cell extracts.
With an eye towards future implementation of this strategy for studies in vivo and in cell extracts, we analyzed our recombinant Fin1 phosphatase reaction products again after mixing them with a large excess of trypsin-digested yeast whole cell extract (Figure 4B). Even within the context of a cell extract, the SRM MS approach was specific enough (Figure 4C-D) that we could faithfully reproduce the relative dephosphorylation rates at all 4 Cdk1 phosphorylation sites (compare 4A and 4B).
In Figure 5 we outline a potential strategy for an in vivo phosphatase assay, using Cdc14 as an example, for which SRM MS could be used to simultaneously monitor a large number of substrate sites under native physiological conditions. The yeast Cdc14 phosphatase is kept inactive during the majority of the cell cycle by its competitive inhibitor, Net1 . Since Cdk sites are preferred Cdc14 substrates, yeast cells can be arrested in early mitosis when Cdk activity is high, for example by chemical treatment with the microtubule depolymerizing agent nocodazole, or by expression of a non-degradable securin (Pds1) mutant . Dephosphorylation of Cdc14 substrates is then “initiated” by addition of galactose to the culture to ectopically express Cdc14 from the GAL1 promoter and aliquots of the culture collected over time. Extracts are made and proteins digested under denaturing conditions  and the phosphorylation status at a large number of specific phosphorylation sites of interest is monitored by LC-SRM MS. SRM methods can be developed based on existing phosphoproteomic data [37-38,42]. This approach would allow for relative rates of dephosphorylation of Cdk sites on many proteins to be determined. We expect that similar types of assays could be designed for a large number of kinases and phosphatases in many different model systems. A current challenge is the ability to detect modified peptides in cell extracts with the required sensitivity. Continued improvements in triple quadrupole instrumentation should make experiments like this more routine in the near future.
This method is not useful for discovery and mapping of phosphorylation sites. The sites must be known or mapped initially by standard MSMS methods. Like other MS analyses of PTMs, success of this method is dependent on generation of detectable peptides by protease digestion. Certain PTMs will inevitably be difficult to detect due to lack of conveniently located digestion sites. Although we have demonstrated previously that two modification sites on a peptide can be individually quantified using SRM MS , existence of several PTMs on a single peptide makes individual quantification difficult. Nonetheless, we conclude from our studies that combining SRM MS with label-free quantification methods provides a powerful, flexible, and accurate approach for the study and characterization of phosphatases and kinases on natural protein substrates. Although we have focused here on reversible protein phosphorylation, this method is equally applicable, in principle, to the study of other PTMs. SRM MS also offers the potential for high throughput analysis of hundreds of modification sites simultaneously. Even with partially purified proteins or large protein complexes, the specificity of SRM is a significant advantage because it largely avoids complications of overlapping signals from other peptides of similar m/z. The use of label-free quantification allows the method to be used with any experimental system, improves sensitivity by eliminating additional chemical reaction or sample processing steps, and obviates the need for costly isotope labeling reagents when monitoring large numbers of PTMs.
1This work was supported in part by National Science Foundation grant MCB 0841748 to MCH and by the Purdue Center for Cancer Research Small Grants Program. This publication was also made possible by support from the Indiana Clinical and Translational Sciences Institute, funded in part by grant # RR 02576 from the National Institutes of Health, National Center for Research Resources. JSM was supported in part by a Purdue Research Foundation graduate assistantship.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.