Robust, Reproducible, and Economical Phosphopeptide Enrichment

Publication Date (Web): December 21, 2018. Copyright © 2018 American Chemical Society. *E-mail: [email protected]. Cite this:J. Proteome Res...
0 downloads 0 Views 1MB Size
Download PDF
Technical Note Cite This: J. Proteome Res. XXXX, XXX, XXX−XXX

pubs.acs.org/jpr

Robust, Reproducible, and Economical Phosphopeptide Enrichment Using Calcium Titanate Adnan Ahmed, Vijay J. Raja,† Paola Cavaliere, and Noah Dephoure* Department of Biochemistry, Sandra and Edward Meyer Cancer Center, Weill Cornell Medical College, New York, New York 10021, United States

Downloaded via IOWA STATE UNIV on January 3, 2019 at 22:42:11 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: Mass-spectrometry-based phosphoproteomics has revolutionized phosphoprotein analysis and enhanced our understanding of diverse and fundamental cellular processes important for human health and disease. Because of their relative scarcity, phosphopeptides must be enriched before analysis. Many different enrichment methods and materials have been described, and many reports have made claims about the advantages of particular materials and methodological variations. We demonstrate an effective and highly reproducible single-step enrichment method using an offthe-shelf preparation of calcium titanate. Using two different cell lines and replicate analysis, we show that our method achieves a purity and depth of analysis comparable or superior to a widely used TiO2-based method at a reduced cost and effort. This method provides a new and immediately available tool for expanding the reach of phosphoproteomic inquiry. KEYWORDS: phosphoproteomics, phosphopeptide enrichment, post-translational modifications, shotgun proteomics, data-dependent analysis

■

INTRODUCTION Protein phosphorylation is the most commonly identified and best studied protein posttranslational modification. It acts as versatile chemical switch that can modulate enzyme activity, protein stability, protein interactions, and subcellular localization. In doing so, it regulates nearly every aspect of cellular growth, homeostasis, and proliferation.1 Its misregulation is implicated in numerous human diseases including cancer, diabetes, and neurodegenerative disorders.2,3 Because of its diverse and essential regulatory roles and because of the clinical success of protein kinase inhibitors, understanding the functions of altered phosphorylation in healthy and disease states is crucial for understanding basic biological processes and for identifying new therapeutic targets and treatments.4,5 The enzymes that control reversible protein phosphorylation, protein kinases and phosphatases, have been intensely studied; more than 1000 original journal articles with “protein kinase” or “protein phosphatase” in the title have been published every year since 1989. Despite these efforts, many of the >500 human protein kinases have no known substrates, and more than a third have poorly or undefined functions.6 We have, in all likelihood, only scratched the surface of our understanding of an immensely complex network of protein phosphorylation whose features have profound effects on human health. With improved quantitative mass-spectrometric methods, including the advent of accurate quantitative multiplexing, enabled by isobaric chemical labeling;7,8 mass-spectrometrybased phosphoproteomic methods now provide a robust experimental platform for the temporal- and condition-specific study of protein phosphorylation.9,10 The provided higher © XXXX American Chemical Society

throughput and quantitative platform enable the broad exploration of the mechanisms and functions of protein phosphorylation and the roles of particular protein kinases and phosphatases under diverse conditions and across many samples. However, these experiments remain costly and time-consuming. Their wider adoption will require new sample preparation and analysis methods. The success of large-scale phosphoproteomics arose from and relies upon methods for enriching relatively scarce phosphopeptides from total protein digests prior to massspectrometric analysis.11−14 This is commonly achieved using immobilized metal affinity chromatography (IMAC) or titanium dioxide (TiO2).15 In recent years, optimized methods using commercial preparations of TiO2 (Titansphere) have become popular due to their reliable performance.16,17 However, to achieve deep coverage of the phosphoproteome, it is necessary to start with large amounts of protein, as much as 10 mg, and even larger amounts of TiO2.9 Analyzing many samples, such as is done in isobaric-labeling-based multiplexing, could require up to 400 mg of commercial TiO2 at a significant cost. In our search for more economical solutions, we were surprised to find very little data examining the relative phosphopeptide enrichment efficacy of different sources and forms of TiO2. Thus we set out to perform such tests, starting with the two primary mineral sources and crystal forms of TiO2, rutile and anatase, and to compare them to the commonly used Titansphere particles from GL Sciences. Received: November 9, 2018 Published: December 21, 2018 A

DOI: 10.1021/acs.jproteome.8b00883 J. Proteome Res. XXXX, XXX, XXX−XXX

Technical Note

Journal of Proteome Research

pelleted with a low speed spin, the flow-through peptides were pipetted off, and the resin was washed twice with 0.5 mL of WBB, followed by three washes with 50% acetonitrile, 0.1% trifluoroacetic acid (TFA). Phosphopeptide enrichment with the TFA-only method was similar to that described for Titansphere TiO2 but used a single buffer (50% ACN, 1% TFA) for all conditioning, peptide resuspension, binding, and wash steps and used only three total washes. Both methods used the same elution protocol. Peptides were eluted in two steps using 200 μL of 2.5% ammonium hydroxide. After a brief vortex and a 10 min incubation, the beads were spun down and the liquid was removed. Both eluates were pooled and dried in a SpeedVac. They were subsequently resuspended in 1% formic acid and desalted on in-house-assembled C18 stage tips19 using six 1.8 mm Empore C18 (3M) discs cut with a 15 G cannula per tip. Peptides were eluted into vial inserts and dried in a SpeedVac. A single replicate of CaTiO3 enrichment from HeLa cells generated a relatively low phosphopeptide enrichment level of 75%. Note that this was manifested as an increase in nonphosphorylated peptides, not a decrease in phosphopeptide identification. This trial generated only slightly fewer unique phosphopeptides than the other two trials and generated more unique phosphopeptides than any of the three Titansphere TiO2 enrichments in HeLa cells. This is consistent with early pilot experiments in which we used a 50:1 ratio of CaTiO3/peptides, where we also observed an increase in the number of nonphosphorylated peptides with no apparent decrease in the number of phosphorylated peptides.

Alongside these different TiO2 preparations, we also included a less commonly used material, calcium titanate (CaTiO3), that was reported to provide high phosphopeptide selectivity.18 Our results demonstrate the efficacy of a simple and economical method for purifying phosphopeptides from complex peptide mixtures using CaTiO3.

■

EXPERIMENTAL SECTION

Cell Culture, Lysis, and Digestion

Jurkat and HeLa cells were grown in RPMI-1640 (Sigma no. R8758) supplemented with 10% fetal bovine serum (FBS) using standard cell culture techniques. Jurkat cells were harvested by gentle centrifugation and washing in phosphatebuffered saline (PBS). HeLa cells were washed in PBS and harvested by scraping. Proteins were harvested and precipitated via addition of 6.5% trichloroacetic acid (TCA) and collected via centrifugation. The pellet was washed with icecold acetone, dried, and then resuspended in 8 M urea, 50 mM ammonium bicarbonate (Ambic). Protein concentration was measured by a Pierce BCA assay (Thermo Scientific no. 23225). Proteins were reduced with the addition of dithiothreitol (DTT) to 5 mM final concentration and incubated at room temperature for 1 h. Iodoacetamide was added to a final concentration of 15 mM, and protein solutions were incubated at room temperature in the dark for 1 h. Iodoacetamide was quenched by the addition of an additional 10 mM equivalent of DTT. Protein solutions were diluted to 2 M urea by the addition of three volumes of 50 mM Ambic. Proteins were digested by the addition of lysyl endopeptidase (lysC) (Wako USA, no. 1202541) at a ratio of 1:125 (enzyme/substrate) with gentle agitation overnight at room temperature. Samples were diluted to 1 M urea with one volume of 50 mM Ambic and digested with a 1:125 ratio of sequencing-grade trypsin (Promega no. V5113) for 6 h at 37 °C with gentle agitation. Peptides were acidified by the addition of neat formic acid (FA) to 1% final concentration. Insoluble material was removed by centrifugation before desalting on tC18 Sep-Pak cartridges (Waters). Peptides were loaded on prewet cartridges in 1% FA, washed with three column volumes of 1% FA, and eluted with 70% acetonitrile (ACN) and 1% FA and dried in a SpeedVac. Purified peptides were reconstituted in HPLC-grade water, and the concentration was measured using the Pierce colorimetric peptide assay kit (Thermo no. 23275) following the vendorsupplied protocol. Peptides were divided for further analysis into 1 mg aliquots and dried down.

Mass-Spectrometric Analysis

Peptides were reconstituted in 5 μL of 5% formic acid. For all samples, we analyzed 2 μL (40%) of the final sample. LC− MS/MS analysis was performed on a Thermo Orbitrap Fusion mass spectrometer equipped with an easy nLC-1000 HPLC (Thermo). Peptides were separated by reverse-phase chromatography on a self-packed 75 μm ID fused silica (polymicro) laser-pulled electrospray emitter (Sutter P2000). The column was packed with 40 cm of 1.8 μm, 120 Å pore size C18 resin (Sepax GP-C18, Sepax Technologies, Newark, DE) behind 0.5 cm of 5 μm ProntoSIL-300-5-C4 (Bischoff Chromatography) using a 75 min gradient of 5−25% buffer B (ACN, 0.1% FA) at a flow rate of 350 nL/min. Peptides were detected using a Top20 method. For each cycle, one full MS scan of m/z = 375−1400 was acquired in the Orbitrap at a resolution of 120 000 with AGC target = 5 × 105. Each full scan was followed by the selection of up to 20 of the most intense ions for collision-induced dissociation (CID) fragmentation and MS/MS analysis in the linear ion trap. Selected ions were excluded from further analysis for 40 s. Ions with charge 1+ or that were unassigned were also rejected. Maximum ion accumulation times were 100 ms for each full MS scan and 35 ms for MS/MS scans. All scans were collected in centroid mode. MS2 spectra were searched using SEQUEST v.28 (rev. 13) against a composite database containing the set of 20 193 reviewed human sequences (UniProt, downloaded March 18, 2016) and its reversed complement with the following parameters: a precursor mass tolerance of ±20 ppm; 1.0dalton product ion mass tolerance; tryptic/lysC digestion (cterminal to Arg or Lys without +1 Pro prohibition); up to two missed cleavages; a static modification of carbamidomethyla-

Phosphopeptide Enrichment

We used the following enrichment materials: anatase TiO2 (Sigma no. 248576), rutile TiO2 (Sigma no. 224227), Titansphere TiO2 (GL Sciences no. 1400B500), and CaTiO3 (Alfa Aesar no. 1397). The cost comparison of CaTiO3 and Titansphere TiO2 was based on the use of 2.5 times more CaTiO3 at $485/500 mg of Titansphere and $95/500 g of CaTiO3. Phosphopeptide enrichment using Titansphere TiO2 particles was performed following a previously published optimized protocol16 that uses 2 M lactic acid in the wash/ binding buffer and a bead-to-peptide ratio of 4:1 (wt/wt). For 1 mg of digested extract, vacuum-dried, desalted peptides were resuspended in 500 μL of wash/binding buffer (WBB = 50% acetonitrile, 2 M lactic acid) with 4 mg of prewashed beads and vortexed at room temperature for 60 min. The beads were B

DOI: 10.1021/acs.jproteome.8b00883 J. Proteome Res. XXXX, XXX, XXX−XXX

Technical Note

Journal of Proteome Research

identified after CaTiO3 and Titansphere TiO2 enrichment are provided as Supplemental Dataset 3. For the comparisons in Figure 2c,d and Figure S-1c,d, we selected the spectral match with the greatest precursor signal-to-noise ratio in each subgrouping. Amino acid frequency analysis was performed on exclusive and shared peptide sequences using nonredundant sequences within each set. Background amino acid frequencies were determined from a set of 14 162 nonredundant nonphosphoenriched peptides from a lysC/Tryptic digest of Jurkat extract identified in a single 85 min run (provided as Supplemental Dataset 4). For both Jurkat and HeLa cell samples, we identified the set of phosphopeptides identified by both CaTiO3 and Titansphere enrichment. Sequences were classified as common to both or exclusive to one of the two methods, and peptide properties were calculated for all unique phosphopeptide sequences and classified into six subsets: Jurkat shared, Jurkat CaTiO3 exclusive, Jurkat Titansphere exclusive, HeLa shared, HeLa CaTiO3 exclusive, and HeLa Titansphere exclusive. The following peptide properties were calculated using the noted regular expression where appropriate: missed cleavages (K.|R^[P]), length, calculated mass, Kyte−Doolittle hydrophobicity,22 the number of charged residues ([DEHKR]), acidic residues ([DE]), basic residues ([HKR]), aromatic residues ([FHWY]), aliphatic residues ([AILV]), small residues ([ACDGNPSTV]), tiny residues ([ACGST]), polar residues ([DEHKNQRST]), and nonpolar residues ([ACFGILMPVWY]). Properties were reported as both a simple per-peptide sum of occurrences and a peptide length-normalized frequency. The distributions of Ser, Thr, and Tyr phosphorylation sites were determined from sites with Ascore ≥19.23 To be included as shared, sites were required to have Ascore ≥19 with both enrichment methods. Because we were attempting to detect phospho-amino acid biases, we limited this analysis to sites that were identified on singly phosphorylated peptides to eliminate any ambiguity about which phospho-amino acid was captured. The distributions of singly, doubly, and triply phosphorylated species were determined from the same set of unique phosphopeptide sequences described in Figure 2b and Figure S-1b. All localized sites (Ascore ≥19) are listed in Supplemental Dataset 5.

tion on cysteine (+57.0214); and dynamic modifications of methionine oxidation (+15.9949) and serine, threonine, and tyrosine phosphorylation (+79.9663). Peptide spectral matches (PSMs) were filtered to 1% false discovery rate (FDR) using the target-decoy strategy20 combined with linear discriminant analysis (LDA)21 using several different parameters including the SEQUEST Xcorr and ΔCn′ scores and precursor mass error. Linear discriminant models were calculated for each LC−MS/ MS run with peptide matches to forward and reversed protein sequences as positive and negative training data. Peptide spectral matches within each run were sorted in descending order by discriminant score and filtered to a 1% FDR, as revealed by the number of decoy sequences remaining in the dataset. The data were further filtered to control protein-level FDRs. Peptides from all runs were combined and assembled into proteins. Protein scores were derived from the product of all LDA peptide probabilities, sorted by rank, and filtered to 1% FDR, as described for peptides. The FDR of the remaining peptides fell markedly after protein filtering. The complete lists of filtered PSMs for the 26 runs described in Figure 1 and the 12 runs described in Figure 2 and Figure S-

Figure 1. Comparing phosphopeptide enrichment materials. The average total number of phosphopeptides identified from two independent purifications using each of the indicated combinations of resin, buffer, and ratio of resin to peptides (wt/wt). LA, lactic acid; TFA, trifluoroacetic acid. Error bars indicate the standard deviation. Shown on the right are the corresponding fractions of all identified spectra matched to phosphorylated peptides. All experiments used 0.5 mg aliquots of lysC and trypsin-digested Jurkat lysate.

Data Availability

Raw mass-spectrometric data for all described experiments have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository with the dataset identifier .

1 are included as Supplemental Datasets 1 and 2, respectively. Label-free intensity measurements and AScore-derived site assignment probabilities are included for the latter 12 runs. Area-proportional Venn diagrams depicting experimental overlaps were generated using eulerr (http://eulerr.co). CaTiO3 and Titansphere comparisons were performed independently for each of the two human cell lines, Jurkat and HeLa. For initial overlap comparisons, we made no assumptions about the correctness of phosphorylation site assignment. Each phosphopeptide species was identified by its amino acid sequence and the number of identified phosphorylation sites. This approach does not attempt to distinguish between same-sequence peptides that differ only in the phosphorylation site assignment and considers any such pair to be identical. This provides a minimum number of unique species and biases toward smaller and more chemically disparate sets of peptides exclusive to the different cell types and enrichment methods. The PSM counts of all exclusive phosphopeptide sequences across Jurkat and HeLa cells and

■

RESULTS AND DISCUSSION We set out to assess the effectiveness of different forms of TiO2 and CaTiO3 for enriching phosphopeptides from complex peptide mixtures and to compare them to an established and widely used method using a commercial preparation of TiO2 with a previously optimized method. For both forms of raw TiO2 and CaTiO3, we used Jurkat cell extracts digested with lysyl endopeptidase (lysC) and trypsin to test two different ratios of enrichment substrate to peptides, 4:1 and 10:1, and two different protocols, the optimized Titansphere protocol, which uses a solution of 2 M lactic acid in 50% ACN, followed by washes in a ACN and TFA buffer, and the other with a simplified buffer system using only a 50% ACN and 1% TFA buffer. We compared these to the optimized protocol using C

DOI: 10.1021/acs.jproteome.8b00883 J. Proteome Res. XXXX, XXX, XXX−XXX

Technical Note

Journal of Proteome Research

Figure 2. Benchmarking CaTiO3 phosphopeptide enrichment against Titansphere TiO2 enrichment in Jurkat Cell extracts. (a) CaTiO3 and Titansphere provide comparable numbers of phosphopeptides with similar purity. The total number of spectra matched to phosphopeptides (blue) and nonphosphorylated (orange) peptides is shown. (b) Overlap analysis of the unique sequences and sites identified by three trials of each method (i), overlap of unique phosphopeptides between the combined datasets for each (ii), overlap of confidently localized sites between methods (iii), and overlap of all spectral matches between methods (iv). (c) Commonly identified sequences are of higher intensity. The distribution of maximum log2 signal intensities of phosphopeptides is plotted separately for shared and exclusive (excl.) phosphopeptides identified by each method. (d) Peptides identified by both CaTiO3 and Titansphere enrichment have similar intensities. For each unique sequence in common, the maximum signal-to-noise value for each method is plotted. (e) Amino acid frequencies of CaTiO3 and Titansphere exclusives phosphopeptides are plotted as the log2 ratio of their frequencies relative to those found in a single 85 min analysis of unenriched HeLa extract peptides. (f) Distribution of Ser, Thr, and Tyr phosphorylations among confidently localized singly phosphorylated peptides exclusive to each method. (g) Titansphere enrichment identifies more multiply phosphorylated peptides. The distributions of singly (1-P), doubly (2-P), and triply (3-P) phosphorylated peptides found with each method independently are shown along with the number of shared sites. Similar data for HeLa-derived peptides appear in Figure S1.

comparison. First, we used a conservative approach in which we grouped peptides with the same amino acid sequence and the same total number of phosphorylation sites to generate the minimum number of distinct phosphopeptides present in the dataset. By this metric, we identified 18 229 unique phosphopeptide sequences from the Jurkat cell extracts, 11 410 of which (63%) were identified by both methods (Figure 2b-ii). The remaining unique sequences were divided almost equally between the two methods. As an alternative means of comparing the overlap of CaTiO3 and TiO2 phosphorylation datasets, we used the AScore algorithm23 to assess the probability of correct site assignment and examined the overlap of singly phosphorylated peptide-derived sites that were localized with ≥99% probability (Ascore ≥19). We identified 9321 confidently localized sites, 5291 of which (57%) were found in both datasets, with the remainder again being split nearly evenly between the two (Figure 2b-iii). Parallel experiments performed with HeLa cells generated highly similar results (Figure S-1). Because we identified many phosphopeptides repeatedly, we also examined the overlap considering the fraction of total spectral matches from common and unique sequences. Notably, the vast majority of matched spectra corresponded to shared sequences. Of the 102 353 PSMs identified collectively in six runs, 95 986 (94%) corresponded to the 11 410 sequences found at least once with both methods (Figure 2b-iv). Shared sequences were identified by an average of 8.4 spectra. Phosphopeptides found exclusively after enrichment with either CaTiO3 or Titansphere were identified at a far lower rate, suggesting that they represent less abundant species. The number of PSM sequences found exclusively in

Titansphere resin. All tested materials and conditions generated sizable phosphopeptide datasets from a single 85 min LC−MS/MS analysis on an Orbitrap Fusion mass spectrometer (Figure 1, Supplemental Dataset 1). CaTiO3 outperformed both rutile and anatase TiO2. Rutile TiO2 was the only material that failed to surpass 5000 identified phosphopeptides under at least one condition. Surprisingly, a crude preparation of −325 mesh (particle size 2000 times less than that for Titansphere TiO2. The analysis of the identified phosphopeptides revealed a large overlap but some clear differences. Phosphopeptide comparisons are complicated by the frequent inability to unambiguously localize the site of phosphorylation to a single amino acid. We used two different methods to make this D

DOI: 10.1021/acs.jproteome.8b00883 J. Proteome Res. XXXX, XXX, XXX−XXX

Technical Note

Journal of Proteome Research Table 1. Peptide Sequence Propertiesa

a Average number or value for each feature per peptide is listed for each indicated subset of the data. mc is missed cleavage, defined by the regular expression “K.|R^[P]”, length is the number of amino acid residues, mw is molecular weight in daltons, and basic is the sum of the HIS, ARG, and LYS. Additional peptide properties appear in Table S-1.

Figure 3. Multiply phosphorylated peptides are longer. (a) The number of phosphorylation sites per peptide is positively correlated with peptide length. The average number of sites per sequence and the distribution of the number of peptides with a given number of sites per length are shown. Data are derived from the full set of 24 446 unique sequence and site count combinations identified across all 12 runs (3× Titansphere + 3× CaTiO3 for each of Jurkat and HeLa cell extracts). (b) The average number of sites per peptide of a given length is shown separately for each of the four cell source and enrichment method variations.

only one of the two experiments was similar, 6210 for CaTiO3 and 5747 for Titansphere. Exclusive sequences were identified by an average of only 1.8 PSMs, indicating that they were usually not found in all three replicates of each method. We further inspected the observed intensity of common and exclusive sequences by looking at the distribution of precursor signal-to-noise measurements. Titansphere and CaTiO 3 exclusive peptide intensities were, on average, three to four times lower than commonly identified ones (Figure 2c and Figure S-1c). We then looked at the set of commonly identified peptides to see if either method generated greater signal intensities that might benefit quantitative analysis. Comparing the maximum signal-to-noise value for each method showed a strong linear correlation (R2 = 0.73) and no intensity bias (Figure 2d). Although both methods produced a comparable quality and depth of phosphopeptide datasets and the vast majority of identified spectra matched sequences found with both, each also generated a sizable number of unique phosphopeptides and phosphorylation sites. To better characterize these differences, we compared a range of peptide properties including length, missed cleavages, mol wt, and hydrophobicity, along with the frequency of different classes of amino acids (Table 1 and Table S-1). Titansphere-enriched peptides were slightly longer, heavier, more likely to have missed cleavages, and more basic. A closer examination of individual amino acid frequencies revealed a marked increase in the frequency of the basic residues histidine and arginine in the Titansphere-exclusive phosphopeptides (Figure 2e, Figure S-1e, and Table 1). The subtle but reproducible increase in

peptide length and mol wt observed with Titansphere can likely be explained by an increased affinity for basic residues whose numbers are increased on missed-cleavage peptides. By looking at confidently localized sites (Ascore ≥19), we found that the relative proportion of phosphoserine and phosphothreonine sites was nearly identical (Figure 2f). Titansphere enrichment recovered slightly more tyrosine phosphorylated sites, 72 (36 exclusive) versus 54 (18 exclusive) for CaTiO3. Although subtle, a similar disparity was also observed in HeLa cells (Figure S-1f), where we identified 105 (45 exclusive) versus 74 (14 exclusive) localized phosphotyrosine sites. The most notable difference between the two methods was that Titansphere enrichment recovered more multiply phosphorylated sequences (Figure 2g). Each method identified a common set of 3664 multiply phosphorylated sequences, but Titansphere identified nearly twice as many sequences that were doubly or triply phosphorylated (1386 versus 755). This result was recapitulated in HeLa cells (Figure S-1g). The increased length of Titansphere-enriched phosphopeptides, likely due to the higher incidence of missed cleavages, may account for the observed increase in multiply phosphorylated peptides. Longer peptides were observed to have more phosphorylation sites in both cell lines and with both methods (Figure 3), and peptides with more phosphorylation sites were longer; doubly and triply modified sequences were on average three and five residues longer than singly phosphorylated peptides, respectively. These data demonstrate how subtle changes in purification methods can generate distinct phosphopeptide datasets. E

DOI: 10.1021/acs.jproteome.8b00883 J. Proteome Res. XXXX, XXX, XXX−XXX

Technical Note

Journal of Proteome Research

■

(2) Chen, M. J.; Dixon, J. E.; Manning, G. Genomics and evolution of protein phosphatases. Sci. Signaling 2017, 10 (474), No. eaag1796. (3) Cohen, P. The role of protein phosphorylation in human health and disease. The Sir Hans Krebs Medal Lecture. Eur. J. Biochem. 2001, 268 (19), 5001−10. (4) Ferguson, F. M.; Gray, N. S. Kinase inhibitors: the road ahead. Nat. Rev. Drug Discovery 2018, 17 (5), 353−377. (5) Cohen, P.; Alessi, D. R. Kinase drug discovery−what’s next in the field? ACS Chem. Biol. 2013, 8 (1), 96−104. (6) Oprea, T. I.; Bologa, C. G.; Brunak, S.; Campbell, A.; Gan, G. N.; Gaulton, A.; Gomez, S. M.; Guha, R.; Hersey, A.; Holmes, J.; Jadhav, A.; Jensen, L. J.; Johnson, G. L.; Karlson, A.; Leach, A. R.; Ma’ayan, A.; Malovannaya, A.; Mani, S.; Mathias, S. L.; McManus, M. T.; Meehan, T. F.; von Mering, C.; Muthas, D.; Nguyen, D. T.; Overington, J. P.; Papadatos, G.; Qin, J.; Reich, C.; Roth, B. L.; Schurer, S. C.; Simeonov, A.; Sklar, L. A.; Southall, N.; Tomita, S.; Tudose, I.; Ursu, O.; Vidovic, D.; Waller, A.; Westergaard, D.; Yang, J. J.; Zahoranszky-Kohalmi, G. Unexplored therapeutic opportunities in the human genome. Nat. Rev. Drug Discovery 2018, 17 (5), 377. (7) McAlister, G. C.; Nusinow, D. P.; Jedrychowski, M. P.; Wuhr, M.; Huttlin, E. L.; Erickson, B. K.; Rad, R.; Haas, W.; Gygi, S. P. MultiNotch MS3 enables accurate, sensitive, and multiplexed detection of differential expression across cancer cell line proteomes. Anal. Chem. 2014, 86 (14), 7150−8. (8) Ting, L.; Rad, R.; Gygi, S. P.; Haas, W. MS3 eliminates ratio distortion in isobaric multiplexed quantitative proteomics. Nat. Methods 2011, 8 (11), 937−40. (9) Erickson, B. K.; Jedrychowski, M. P.; McAlister, G. C.; Everley, R. A.; Kunz, R.; Gygi, S. P. Evaluating multiplexed quantitative phosphopeptide analysis on a hybrid quadrupole mass filter/linear ion trap/orbitrap mass spectrometer. Anal. Chem. 2015, 87 (2), 1241−9. (10) Huang, F. K.; Zhang, G.; Lawlor, K.; Nazarian, A.; Philip, J.; Tempst, P.; Dephoure, N.; Neubert, T. A. Deep Coverage of Global Protein Expression and Phosphorylation in Breast Tumor Cell Lines Using TMT 10-plex Isobaric Labeling. J. Proteome Res. 2017, 16 (3), 1121−1132. (11) Olsen, J. V.; Blagoev, B.; Gnad, F.; Macek, B.; Kumar, C.; Mortensen, P.; Mann, M. Global, in vivo, and site-specific phosphorylation dynamics in signaling networks. Cell 2006, 127 (3), 635−48. (12) Villen, J.; Beausoleil, S. A.; Gerber, S. A.; Gygi, S. P. Large-scale phosphorylation analysis of mouse liver. Proc. Natl. Acad. Sci. U. S. A. 2007, 104 (5), 1488−93. (13) Ficarro, S. B.; McCleland, M. L.; Stukenberg, P. T.; Burke, D. J.; Ross, M. M.; Shabanowitz, J.; Hunt, D. F.; White, F. M. Phosphoproteome analysis by mass spectrometry and its application to Saccharomyces cerevisiae. Nat. Biotechnol. 2002, 20 (3), 301−5. (14) Posewitz, M. C.; Tempst, P. Immobilized gallium(III) affinity chromatography of phosphopeptides. Anal. Chem. 1999, 71 (14), 2883−92. (15) Leitner, A. Enrichment Strategies in Phosphoproteomics. Methods Mol. Biol. 2016, 1355, 105−21. (16) Kettenbach, A. N.; Gerber, S. A. Rapid and reproducible singlestage phosphopeptide enrichment of complex peptide mixtures: application to general and phosphotyrosine-specific phosphoproteomics experiments. Anal. Chem. 2011, 83 (20), 7635−44. (17) Fukuda, I.; Hirabayashi-Ishioka, Y.; Sakikawa, I.; Ota, T.; Yokoyama, M.; Uchiumi, T.; Morita, A. Optimization of enrichment conditions on TiO2 chromatography using glycerol as an additive reagent for effective phosphoproteomic analysis. J. Proteome Res. 2013, 12 (12), 5587−97. (18) Li, X. S.; Chen, X.; Sun, H.; Yuan, B. F.; Feng, Y. Q. Perovskite for the highly selective enrichment of phosphopeptides. J. Chromatogr A 2015, 1376, 143−8. (19) Rappsilber, J.; Ishihama, Y.; Mann, M. Stop and go extraction tips for matrix-assisted laser desorption/ionization, nanoelectrospray, and LC/MS sample pretreatment in proteomics. Anal. Chem. 2003, 75 (3), 663−70.

CONCLUSIONS We have demonstrated a remarkably robust and scalable method for generating large pools of enriched phosphopeptides for mass-spectrometric analysis using widely available and affordable CaTiO3. In experiments in two cell lines, we showed it to be as reliable and effective as a popular method using a commercial preparation of TiO2. Whether used to supplant or to supplement existing methods, the ease and economy of CaTiO3 phosphopeptide enrichment will facilitate future largescale phosphoproteomic analyses of diverse samples.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jproteome.8b00883.

■

Figure S-1. HeLa-cell-derived phosphopeptide comparisons of CaTiO3 and Titansphere enrichment. Table S-1. Expanded phosphopeptide properties (PDF) Dataset 1. Peptide identifications of all filtered PSMs identified with the different enrichment resins and protocols. Dataset 2. All filtered PSMs from triplicate analysis of each of the Jurkat and HeLa cell extracts by CaTiO 3 and Titansphere including AScores and intensity measurements. Dataset 3. Per sample accounting of all unique phosphopeptide sequences in Dataset 2. Dataset 4. All unique peptide sequences from an analysis of non-phospho-enriched Jurkat cell extract digest. Dataset 5. All confidently localized phosphorylation sites and the samples in which they were identified (XLSX)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Noah Dephoure: 0000-0001-5802-2657 Present Address †

V.J.R.: BioMarin Pharmaceutical, Inc., Novato, CA.

Author Contributions

N.D. supervised all aspects of the project. A.A., V.J.R., and P.C. performed experiments. A.A., V.J.R., P.C., and N.D. performed data analysis. N.D. wrote the manuscript. All authors discussed the results and commented on the manuscript. Notes

The authors declare no competing financial interest. Raw mass-spectrometric data have been deposited in the PRIDE archive (https://www.ebi.ac.uk/pride/archive/) as dataset PXD011145.

■

ACKNOWLEDGMENTS We thank Yogindrya Vedyvas and Moonsoo Jin for providing and verifying the identify of Jurkat and HeLa cells. N.D. and V.J.R. were supported in part by a grant from Project ALS.

■

REFERENCES

(1) Pawson, T.; Scott, J. D. Protein phosphorylation in signaling−50 years and counting. Trends Biochem. Sci. 2005, 30 (6), 286−90. F

DOI: 10.1021/acs.jproteome.8b00883 J. Proteome Res. XXXX, XXX, XXX−XXX

Technical Note

Journal of Proteome Research (20) Elias, J. E.; Gygi, S. P. Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry. Nat. Methods 2007, 4 (3), 207−14. (21) Huttlin, E. L.; Jedrychowski, M. P.; Elias, J. E.; Goswami, T.; Rad, R.; Beausoleil, S. A.; Villen, J.; Haas, W.; Sowa, M. E.; Gygi, S. P. A tissue-specific atlas of mouse protein phosphorylation and expression. Cell 2010, 143 (7), 1174−89. (22) Kyte, J.; Doolittle, R. F. A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 1982, 157 (1), 105− 32. (23) Beausoleil, S. A.; Villen, J.; Gerber, S. A.; Rush, J.; Gygi, S. P. A probability-based approach for high-throughput protein phosphorylation analysis and site localization. Nat. Biotechnol. 2006, 24 (10), 1285−92.

G

DOI: 10.1021/acs.jproteome.8b00883 J. Proteome Res. XXXX, XXX, XXX−XXX