This protocol describes a procedure to extract and enrich phosphopeptides from prostate cancer cell lines or tissues for an analysis of the phosphoproteome via mass spectrometry-based proteomics.
Phosphoproteomics involves the large-scale study of phosphorylated proteins. Protein phosphorylation is a critical step in many signal transduction pathways and is tightly regulated by kinases and phosphatases. Therefore, characterizing the phosphoproteome may provide insights into identifying novel targets and biomarkers for oncologic therapy. Mass spectrometry provides a way to globally detect and quantify thousands of unique phosphorylation events. However, phosphopeptides are much less abundant than non-phosphopeptides, making biochemical analysis more challenging. To overcome this limitation, methods to enrich phosphopeptides prior to the mass spectrometry analysis are required. We describe a procedure to extract and digest proteins from tissue to yield peptides, followed by an enrichment for phosphotyrosine (pY) and phosphoserine/threonine (pST) peptides using an antibody-based and/or titanium dioxide (TiO2)-based enrichment method. After the sample preparation and mass spectrometry, we subsequently identify and quantify phosphopeptides using liquid chromatography-mass spectrometry and analysis software.
An estimated 165,000 new cases and approximately 29,000 deaths will occur in 2018 due to prostate cancer, representing the most common cancer and second leading cause of cancer-related death in men in the United States1. Early stages of prostate cancer are treatable with resection or radiation therapy of organ-confined disease, where the ten-year recurrence rate is between 20% and 40% for patients who undergo prostatectomy and between 30% and 50% for patients who receive radiation therapy2. Because prostate cancer relies on androgen signaling for growth, surgical and chemical castration therapies are also employed for high-risk patients. However, relapse occurs when the cancer no longer responds to androgen deprivation therapy as evidenced by biochemical recurrence, where the prostate-specific antigen in serum rises again. At this point in the progression, metastases are often detected as well. This advanced stage, called metastatic castration-resistant prostate cancer, represents the lethal form of the disease where the prognosis is a median survival time of less than two years3. Few treatment options are available in late-stage disease, including second-generation antiandrogens such as enzalutamide and abiraterone, as well as taxane-based chemotherapy like docetaxel. Despite available treatments, the disease often progresses. Therefore, the discovery and development of novel treatment modalities are necessary to improve the care of prostate cancer patients with advanced disease.
Mass spectrometry (MS)-based approaches provide a global analysis of the proteome through the detection of hundreds to thousands of peptide analytes4. In particular, discovery proteomics, also known as data-dependent acquisition (DDA), can yield the identification and quantitation of thousands of peptides4,5. MS-based discovery proteomics can be further delineated into top-down proteomics, where intact proteins are characterized, and bottom-up (also known as shotgun) proteomics, where peptides are analyzed to characterize proteins5. Thus, in shotgun proteomics, a proteolysis step takes place in the sample preparation preceding the MS analysis to cleave proteins into peptides. At the end, a database search is performed to map the peptides back to the proteins for identification. Label-free as well as several isotope-labeling [e.g., stable isotope labeling by amino acids in cell culture (SILAC)] methods can be used to quantitatively compare peptides between samples6,7. While isotope labeling techniques are the gold standard, label-free methods have demonstrated similar quantification accuracies8,9 and have comparable tradeoffs between sensitivity and specificity10. Label-free quantitation provides greater coverage and permits comparisons between many more samples, whereas label-based methods are limited by cost and multiplexing capacities6,7,8.
Furthermore, shotgun MS can be also used to interrogate post-translational modifications (PTMs) such as phosphorylation11. Due to the lower stoichiometric nature of phosphopeptides compared to total peptides, several methods are employed to enrich for phosphopeptides, including antibody-based immunoprecipitation of phosphotyrosine (pY) peptides, titanium dioxide (TiO2), and immobilized metal affinity chromatography (IMAC)5,12. Because protein phosphorylation is a key step in many cell signaling pathways, shotgun phosphoproteomics allows researchers to investigate cell signaling changes in different cancers, including breast13, prostate14, renal15, and ovarian,16,17 to better understand cancer biology and to identify potential new targets for therapy.
This label-free shotgun phosphoproteomic method was built and refined based on previous work by the Graeber group18,19,20. This protocol begins by describing the extraction and digestion of proteins and phosphoproteins from tissue into peptides. We then detail the enrichment of pY peptides using specific phosphotyrosine antibodies and TiO2. We also describe the enrichment of phosphoserine/threonine (pST) peptides using strong cation exchange (SCX) followed by TiO2. This protocol concludes with the submission of samples to an MS facility and the use of MS analysis software to identify and quantify phosphopeptides and their corresponding phosphoproteins. The application of this protocol can extend beyond prostate cancer into other cancers and fields outside of oncology.
Experiments using xenograft tumors were approved by the Rutgers University Institutional Animal Care and Use Committee as set forth under the guidelines of the National Institutes of Health.
1. Protein Extraction
2. Lysate Digestion
3. Reverse Phase Extraction
4. Immunoprecipitation and Enrichment of pY Peptides24
5. Titanium Dioxide Enrichment25 of pY Peptides
6. Desalting pY Peptides for MS Analyses
7. Reverse Phase Extraction of pST Peptides
8. Strong Cation Exchange (SCX) of pST Peptides
9. Titanium Dioxide Enrichment of pST Peptides
10. Desalting pST Peptides for MS Analyses
11. Mass Spectrometry Analysis
This protocol describes in detail a method for protein extraction and digestion followed by phosphopeptide enrichment and subsequent MS analysis (Figure 1). The compositions of all the buffers and solutions that are used in this protocol are listed in Table 1. The sequential use of Lys-C and trypsin provides an efficient digestion. A Coomassie-stained gel of pre-digested lysate confirms the presence of proteins, while staining of post-digested lysate confirms the complete digestion (Figure 2A). For a complete digestion, no bands should appear above 15 kDa, except the 30 kDa and 23.3 kDa bands for Lys-C and trypsin, respectively. The addition of Lys-C also reduces the number of missed cleavages (Figure 2B). Because pY peptides represent only 2% of the phosphoproteome28, immunoprecipitation of the pY peptides using a pY-specific antibody is the first step of pY peptide enrichment. The resulting supernatant becomes the input for pST peptide enrichment. The pY immunoprecipitation effectively separates pY peptides from pST peptides where on average 85% of the phosphopeptides identified from the pY preparation are pY (Figure 3A) and over 99% of the phosphopeptides identified from the pST preparation are pST (Figure 3B). Titanium dioxide is used to enrich for phosphopeptides in both preparations. The expected percentage of peptides in the MS-ready preparation that are phosphorylated is between 30 – 50% (Figure 4A). The variability in the phosphopeptide enrichment percentage may be greater in the pY preparation as a result of there being many fewer pY peptides than pST peptides. In terms of phosphopeptide species, the majority of the phosphopeptides detected have a single or double phosphoryl group (Figure 4B).
After performing mass spectrometry, the MS raw files are loaded into an MS analysis software. The parameter settings used in the experiment are listed in Table 3 but will vary from software to software and may vary from version to version. The parameters that are not listed were left as default, including an FDR cutoff of 1% for peptide-spectrum matching (PSM) with a minimum Andromeda score of 40 for the identification of modified peptides27. Setting a localization probability cutoff of greater than 0.75 filters out approximately 5% of the pY peptides and 15% and 34% of the pS and pT peptides, respectively (Figure 5A). After applying these filters, the expected number of phosphopeptide identifications at the end of the MS analysis is approximately 300 pY peptides (for 5 mg of the starting protein) and about 7,500 pS peptides and 640 pT peptides (for 2.5 mg of the starting peptide amount) from the respective enrichment preparations (Figure 5B). The number of replicates and the variability of the phosphopeptide signal intensity determines adequate powering for statistical comparisons. In four separate experiments with groups containing either biological duplicates or triplicates, the percent coefficients of variation (%CV) for detected phosphopeptides were calculated. Distributions of lower variability (e.g., pST groups 1 – 5 in Figure 5C) indicate that the sample collection, preparation, and mass spectrometry runs were consistent. On the other hand, distributions of higher variability (e.g., pST group 6 in Figure 5C) indicates noisier data that would require larger fold-changes to detect significant differences in downstream differential analyses.
Figure 1: Workflow diagram. Proteins from samples are extracted and digested. Peptides are extracted by solid phase extraction (SPE), and phosphotyrosine (pY) peptides are immunoprecipitated. In parallel, the phosphoserine/threonine (pST) peptides are enriched from the supernatant in the pY immunoprecipitation step. Strong cation exchange (SCX) is performed on the supernatant to remove highly charged peptides to reduce the ion suppression12. Both preparations undergo phosphopeptide enrichment via titanium dioxide (TiO2). After sample cleanup, liquid chromatography-tandem mass spectrometry (LC-MS/MS) is performed to measure the phosphopeptide abundance. The raw data is then loaded into an MS analysis software to identify phosphopeptides. Please click here to view a larger version of this figure.
Figure 2: Evaluation of digestion. (A) Three samples with 12.5 µg of lysate pre-digestion, post-Lys-C digestion, and post-trypsin digestion are shown. A Coomassie gel-stain test shows a clean digestion after sequential use of Lys-C and trypsin. The molecular weight (MW) size markers are in kilodaltons (kDa). (B) A reduction in missed cleavages is observed after Lys-C was added to the protocol. The percentage of phosphopeptides without missed cleavages increased from 48% to 64% and from 60% to 84% on average for pY and pST enrichment preparations, respectively. The graphs summarize the data obtained from two experiments performed without Lys-C and five experiments performed with Lys-C. The error bars are standard deviations representing 38 pY and 38 pST samples from 2 separate experiments (without Lys-C) and 62 pY and 60 pST samples from 5 separate experiments (with Lys-C). Please click here to view a larger version of this figure.
Figure 3: Enrichment of pY and pST phosphopeptides. These panels show the percentages of pSTY phosphopeptides from either (A) the pY or (B) the pST enrichment preparations. The pY enrichment by pY immunoprecipitation and titanium dioxide resulted in 85% phosphopeptides being for pY peptides, while only 0.1% of the phosphopeptides in the pST enrichment are pY. These values were drawn from examining the Phospho (STY)Sites.txt file of one representative experiment after filtering out contaminants, reverse sequences, and phosphopeptides with localization probabilities less than 0.75. Please click here to view a larger version of this figure.
Figure 4: Phosphopeptide enrichment with titanium dioxide. (A) The percentage of detected phosphopeptides (relative to total peptides) from samples in four separate experiments is shown. (B) This panel shows the average composition of mono-, double-, and multi-phosphorylated peptides in four separate experiments. The error bars in panel A are standard deviations. Please click here to view a larger version of this figure.
Figure 5: Expected phosphoresidue identifications. (A) This panel shows the phosphorylation localization probabilities of IDs from pY enrichment (left) and pST enrichment (right). The mean percentage of IDs that meet the > 0.75 probability cutoff is 93%, 75%, and 52% for pY, pS, and pT, respectively. (B) The mean number of IDs with a >0.75 localization probability is 300 for pY, 7,500 for pS, and 640 for pT. (C) This panel shows violin plots of the percent coefficient of variation (%CV) of the phosphopeptides. An evaluation of %CV was only performed if a signal intensity value was detected in each biological replicate or triplicate group. Data was taken from four separate experiments. The error bars in panels A and B are standard deviations from 34 pY and 34 pST samples from 4 separate experiments. Please click here to view a larger version of this figure.
Buffer | Volume | Composition | |
6 M guanidinium chloride lysis buffer | 50 mL | 6 M guanidinium chloride, 100 mM tris pH 8.5, 10 mM tris (2-carboxyethyl) phosphine, 40 mM chloroacetamide, 2 mM sodium orthovanadate, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 500 mg n-octyl-glycoside, ultra-pure water to volume | |
100 mM sodium pyrophosphate | 50 mL | 2.23 g sodium pyrophosphate decahydrate, ultra-pure water to volume | |
1M β-glycerophosphate | 50 mL | 15.31 g β-glycerophosphate, ultra-pure water to volume | |
5% trifluoroacetic acid | 20 mL | Add 1 mL of 100% trifluoroacetic acid into 19 mL ultra-pure water | |
0.1% trifluoroacetic acid | 250 mL | Add 5 mL 5% trifluoroacetic acid to 245 mL ultra-pure water | |
pY elution buffer | 250 mL | 0.1% trifluoroacetic acid, 40% acetonitrile, ultra-pure water to volume | |
pST elution buffer | 250 mL | 0.1% trifluoroacetic acid, 50% acetonitrile, ultra-pure water to volume | |
IP binding buffer | 200 mL | 50 mM tris pH 7.4, 50 mM sodium chloride, ultra-pure water to volume | |
25 mM ammonium bicarbonate, pH 7.5 | 10 mL | Dissolve 19.7 mg into 10 mL sterile ultra-pure water, pH to 7.5 with 1 N hydrochloric acid (~10-15 µL/10 ml solution), make fresh | |
1M phosphate buffer, pH 7 | 1,000 mL | 423 mL 1 M sodium dihydrogen phosphate, 577 mL 1 M sodium hydrogen phosphate | |
Equilibration buffer | 14 mL | 6.3 mL acetonitrile, 280 µL 5% trifluoroacetic acid, 1740 µL lactic acid, 5.68 mL ultra-pure water | |
Rinsing buffer | 20 mL | 9 mL acetonitrile, 400 µL 5% trifluoroacetic acid, 10.6 mL ultra-pure water | |
Mass spectrometry solution | 10 mL | 500 µL acetonitrile, 200 µL 5% trifluoroacetic acid, 9.3 mL ultra-pure water | |
Buffer A | 250 mL | 5 mM monopotassium phosphate (pH 2.65), 30% acetonitrile, 5 mM potassium chloride,ultra-pure water to volume | |
Buffer B | 250 mL | 5 mM monopotassium phosphate (pH 2.65), 30% acetonitrile, 350 mM potassium chloride, ultra-pure water to volume | |
0.9% ammonium hydroxide | 10 mL | 300 μL 29.42% ammonium hydroxide, 9.7 mL ultra-pure water |
Table 1: Buffers and solutions. This table shows the compositions of the buffers and solutions used in this protocol.
LC-MS/MS Settings | ||
Parameter | pY Setting | pST Setting |
Sample loading (µL) | 5 | |
Loading flow rate (µL/min) | 5 | |
Gradient flow rate (nL/min) | 300 | |
Linear gradient (percentage 0.16% formic acid, 80% ACN in 0.2% formic acid) | 4 – 15% for 5 min | 4 – 15% for 30 min |
15 – 50% for 40 min | 15 – 25% for 40 min | |
50 – 90% for 5 min | 25 – 50% for 44 min | |
50 – 90% for 11 min | ||
Full scan resolution | 120,000 | |
Number of most intense ions selected | 20 | |
Relative collision energy (%) (HCD) | 27 | |
Dynamic exclusion (s) | 20 |
Table 2: LC-MS settings. This is an example of LC-MS settings in a typical shotgun phosphoproteomic experiment. The samples were loaded on to a trap column. The trap was brought in-line with an analytical column. These settings were optimized for using the LC-MS system listed in the Table of Materials and Reagents. These settings would need to be adjusted for other LC-MS systems.
MaxQuant Parameter Settings | ||
Setting | Action | |
Group-Specific Parameters | ||
Type | Type | Select Standard |
Multiplicity | Set to 1 | |
Digestion Mode | Enzyme | Select Trypsin/P |
Max. missed cleavages | Set to 2 | |
Modifications | Variable modifications | Add Phospho (STY) |
Label-free quantification | Label-free quantification | Select LFQ |
LFQ min. ratio count | Set to 1 | |
Fast LFQ | Check off | |
Miscellaneous | Re-quantify | Check off |
Global Parameters | ||
Sequences | FASTA files | Select fasta file downloaded from UniProt |
Fixed modifications | Add Carbamidomethyl (C) | |
Adv. Identification | Match between runs | Check off |
Match time window | Set to 5 min | |
Alignment time window | Set to 20 min | |
Match unidentified features | Check off | |
Protein quantification | Min. ratio count | Set to 1 |
Folder locations | Modify accordingly |
Table 3: MS analysis software settings. In MaxQuant, the group-specific and global parameters in this table were selected or adjusted. All other parameters remained at default. These experiments were conducted using version 1.5.3.30. The parameters may vary from version to version and from software to software.
Before utilizing this protocol to enrich for phosphopeptides, a careful consideration of the experimental design is critical. Using biological replicates is a more cost-effective use of mass spectrometry resources than technical replicates. The number of replicates that are necessary will depend in part on the variability of the data. A recent study demonstrated that, while increasing the number of replicates beyond three only marginally increases the number of identifications, the number of significant identifications between groups increases with more replicates10.
Due to the lower abundance of phosphoproteins in the cell, sufficient starting protein amounts are necessary to obtain a global phosphoproteome from prostate cancer samples in discovery mode. In these experiments, 5 mg of protein was used. Approximately five nearly confluent 15-cm dishes of cells provide enough protein as input into this protocol, although this will be cell line-dependent. As for tumor tissue, the expected yield of protein is about 6 – 8% of tissue weight. In the in vitro setting, a positive control sample to consider is the addition of 1 mM vanadate for 30 min before harvesting the cells. Vanadate, a competitive protein phosphotyrosyl phosphatase inhibitor, will preserve the tyrosine phosphorylation, thus increasing the number of pY peptide identifications29.
Clean digestion is a key step to maximize phosphopeptide identification. In addition to the Coomassie stain test, the percent of missed cleavages in the data can be used to evaluate digestion efficiency (Figure 2). Quality-control software is available that analyzes missed cleavages and other metrics to assess MS data quality30. While trypsin is the most common, alternative proteases are available5 to address coverage gaps in the proteome where optimal tryptic peptides cannot be generated31. The settings of the MS analysis software would then need to be modified accordingly to adjust for changes in proteases.
The protocol employs immunoprecipitation (for pY enrichment) as well as titanium dioxide (TiO2) to enrich for phosphopeptides. Alternative approaches to enrich for peptides include immobilized metal affinity chromatography (IMAC), other metal oxides for metal oxide affinity chromatography (MOAC) such as aluminum hydroxide, and polymer-based metal ion affinity capture (PolyMAC)5,12. Previous studies have shown that different enrichment methods enrich for different populations of phosphopeptides32. For instance, IMAC enriches more multi-phosphorylated peptides while MOAC preferably enriches for mono-phosphorylated peptides33. The Representative Results of this protocol reflect this observation (Figure 4B). A recent publication demonstrated that combining IMAC and MOAC using a hybrid material could potentially provide greater coverage of phosphopeptide species34. Thus, this protocol could be modified to utilize other enrichment methods in parallel to allow for even more comprehensive phosphoproteomic analyses.
The MaxQuant26 software suite is used to analyze the MS data in this protocol, but commercial applications35 are also available for phosphopeptide identification and quantification. For phosphopeptide identification, a localization probability cutoff is applied. This filter is performed to select for phosphopeptides with a high confidence (i.e., greater than 0.75) in phosphoresidue identification10,28. In other words, the summed probability of all other residues that could potentially contain the phospho-group is less than 0.25. This cutoff could be raised to increase the stringency of the phosphopeptide selection. In regard to the number of identifications, the expected number of pY peptides is in the hundreds, while the expected number of pST peptides is in the high thousands. These values reflect previously observed phosphoproteome distribution where about 2%, 12%, and 86% of the phosphosites are pY, pT, and pS, respectively28.
If the pY and pST enrichment steps are performed in parallel, the sample preparation steps in the protocol can be completed in six days. By pairing with the powerful tool of MS, phosphopeptide enrichment protocols such as this provide a global approach for scientists to collect data to analyze the phosphoproteome in their respective research fields.
The authors have nothing to disclose.
We thank members of the Drake lab for providing advice and input on the manuscript. We also thank the members of the Biological Mass Spectrometry Facility of Robert Wood Johnson Medical School and Rutgers, The State University of New Jersey, for providing advice and performing mass spectrometry on our samples. Larry C. Cheng is supported by the National Institute of General Medical Sciences of the National Institutes of Health under award number T32 GM008339. Thomas G. Graeber is supported by the NCI/NIH (SPORE in Prostate Cancer P50 CA092131; P01 CA168585) and an American Cancer Society Research Scholar Award (RSG-12-257-01-TBE). Justin M. Drake is supported by the Department of Defense Prostate Cancer Research Program W81XWH-15-1-0236, Prostate Cancer Foundation Young Investigator Award, the New Jersey Health Foundation, and a Precision Medicine Initiative Pilot Award from the Rutgers Cancer Institute of New Jersey.
Ultra-Low Temperature Freezer | Panasonic | MDF-U76V | |
Freezer -20 °C | VWR | scpmf-2020 | |
Swing rotor bucket | ThermoFisher Scientific | 75004377 | |
Vacuum manifold | Restek | 26080 | |
Lyophilizer | Labconco | 7420020 | |
CentriVap Benchtop Vacuum Concentrator | Labconco | 7810010 | |
End-over-end rotator | ThermoFisher Scientific | 415110Q | |
Razor blade | Fisher Scientific | 620177 | |
Amicon Ultra-15 Centrifugal Filter Units | Millipore Sigma | UFC901024 | |
Glass culture tubes | Fisher Scientific | 14-961-26 | |
Parafilm | Fisher Scientific | 13-374-12 | |
20G needle | BD | B305175 | |
Kimwipes | Fisher Scientific | 06-666A | |
Screw cap cryotube | ThermoFisher Scientific | 379189 | |
Nunc 15 mL conical tubes | ThermoFisher Scientific | 12-565-268 | |
Gel loading tips | Fisher Scientific | 02-707-181 | |
Millipore 0.2 µm spin filter | Millipore Sigma | UFC30GVNB | |
Low protein-binding Eppendorf tubes | Eppendorf | 22431081 | |
anti-Phosphotyrosine, Agarose, Clone: 4G10 | Millipore Sigma | 16101 | |
27B10.4 antibody | Cytoskeleton | APY03-beads | |
Peptide assay kit | Thermo Scientific | 23275 | Step 7 |
TopTip | PolyLC Inc | TT200TIO.96 | Steps 5 and 9 |
SCX columns (PolySULFOETHYL A) | PolyLC Inc | SPESE1203 | |
3 mL syringe | BD | 309657 | |
Trifluoracetic Acid (TFA) | Fisher Scientific | PI-28904 | |
Acetonitrile (ACN) | Fisher Scientific | A21-1 | |
Lactic acid | Sigma-Aldrich | 69785-250ML | |
Ammonium Hydroxide | Fisher Scientific | A669S-500 | |
Potassium Phosphate Monobasic | Fisher Scientific | BP362-500 | |
Potassium Chloride | Fisher Scientific | BP366-500 | |
Calcium Chloride Dihydrate | Fisher Scientific | BP510-500 | |
Tris Base | Fisher Scientific | BP152-5 | |
Trypsin, TPCK Treated | Worthington Biochemicals | LS003740 | |
Lysyl Endopeptidase | Wako Pure Chemical Industries, Ltd. | 125-05061 | |
MonoTip | PolyLC Inc | TT200TIO.96 | Step 10 |
ZipTip | MilliporeSigma | ZTC18S096 | Step 6 |
nanoEase, MZ peptide BEH C18, 130A, 1.7 μm, 75 μm x 20 cm | Waters | 186008794 | Step 11: analytical column |
Acclaim PepMap 100 C18 LC Columns | ThermoFisher Scientific | 164535 | Step 11: trap column |
Ultimate 3000 RLSCnano System | Dionex | ULTIM3000RSLCNANO | Step 11 |
Q Exactive HF | ThermoFisher Scientific | IQLAAEGAAPFALGMBFZ | Step 11 |
MilliQ water | deionized water used to prepare all solutions and bufferes | ||
Sonic Dismembrator | Fisher Scientific | FB-120 | sonicator |
Polytron System PT | Kinematica AG | PT 10-35 GT | homogenizer |