Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Feb 25.
Published in final edited form as: Anal Chem. 2015 Aug 11;87(17):8837–8844. doi: 10.1021/acs.analchem.5b01833

Comparing Multi-Step IMAC and Multi-Step TiO2 Methods for Phosphopeptide Enrichment

Xiaoshan Yue 1, Alissa Schunter 1, Amanda B Hummon 1,*
PMCID: PMC4766865  NIHMSID: NIHMS760470  PMID: 26237447

Abstract

Phosphopeptide enrichment from complicated peptide mixtures is an essential step for mass spectrometry-based phosphoproteomic studies to reduce sample complexity and ionization suppression effects. Typical methods for enriching phosphopeptides include immobilized metal affinity chromatography (IMAC) or titanium dioxide (TiO2) beads, which have selective affinity and can interact with phosphopeptides. In this study, the IMAC enrichment method was compared with the TiO2 enrichment method, using a multi-step enrichment strategy from whole cell lysate, to evaluate their abilities to enrich for different types of phosphopeptides. The peptide-to-beads ratios were optimized for both IMAC and TiO2 beads. Both IMAC and TiO2 enrichments were performed for three rounds to enable the maximum extraction of phosphopeptides from the whole cell lysates. The phosphopeptides that are unique to IMAC enrichment, unique to TiO2 enrichment, and identified with both IMAC and TiO2 enrichment were analyzed for their characteristics. Both IMAC and TiO2 enriched similar amounts of phosphopeptides with comparable enrichment efficiency. However, phosphopeptides that are unique to IMAC enrichment showed a higher percentage of multi-phosphopeptides, as well as a higher percentage of longer, basic, and hydrophilic phosphopeptides. Also, the IMAC and TiO2 procedures clearly enriched phosphopeptides with different motifs. Finally, further enriching with two rounds of TiO2 from the supernatant after IMAC enrichment, or further enriching with two rounds of IMAC from the supernatant TiO2 enrichment does not fully recover the phosphopeptides that are not identified with the corresponding multi-step enrichment.

Keywords: Phosphoproteomics, multi-step IMAC enrichment, multi-step TiO2 enrichment, phosphopeptides, phosphorylation motifs

INTRODUCTION

Phosphopeptide enrichment is a key step in phosphoproteomic studies. While global phosphoproteomic research is generating widespread interest with the promise of providing a glimpse of phosphorylation events occurring during cellular processes, global phosphoproteomic research is limited by the low abundance of phosphoproteins in the whole cell lysate as well as the high variety of their phosphorylation states 1-3. Phosphopeptides need to be enriched prior to mass spectrometric analysis to reduce the sample complexity and minimize the repression effects from non-phosphorylated peptides 4-6. There are several approaches for enriching phosphopeptides from the peptide mixtures, including immobilized metal affinity chromatography (IMAC) with Fe3+ or Ti4+ 7,8,9, metal oxide affinity chromatography (MOAC) with TiO2 10,11, and immuno-affinity based methods 12. Other methods such as polymer-based metal ion affinity capture (PolyMAC) 13,14 and polymer-oxotitanium-modified gold wafer (Au-P-oxoTi) 15 result in successful phosphopeptide enrichment.

IMAC has a high affinity for phosphopeptides through the attraction of metal cations and negatively charged phosphate groups. On the other hand, MOAC enriches phosphopeptides mainly based on metal atoms in the MOAC resin with high affinity to the oxygen in the phosphate groups. For the two most widely used phosphopeptide enrichment methods, Fe3+-IMAC and TiO2, previous research indicates that IMAC and TiO2 tend to enrich distinct groups of phosphopeptides, and that they have different performance and efficiency in enriching phosphopeptides16-18. However, for large-scale phosphoproteomic studies, the exact differences between the peptide groups that are enriched by IMAC or TiO2 are still unclear. It would be helpful to researchers to select the enrichment material best suited to their sample type, for example, when studying a pathway where particularly acidophilic or basophilic kinase motifs are expected to predominate. Also, it is of interest to investigate whether adding IMAC enrichment steps following TiO2 or TiO2 following IMAC can increase total phosphopeptide recovery and improve the overlap between the two methods.

In our previous research, we reported that multiple rounds of IMAC treatment with optimized conditions enriched similar numbers of phosphopeptides from whole cell lysates as an SCX-IMAC workflow, with one-fourth of the analysis time and one-fifth of the starting material 19. Further fractionating the three enriched pools by high-pH reverse phase increases identifications and eliminates a desalting step prior to LC-MS analysis. In this study, we compared the IMAC enrichment method with the TiO2 enrichment method by utilizing the multi-step enrichment strategy from whole cell lysates and evaluated the efficiency of both enrichment methods. We identified similar amounts of phosphopeptides with both multi-step IMAC and multi-step TiO2 enrichment. The identified phosphopeptides were further divided into three subgroups including 1) those that are unique to IMAC enrichment, 2) those that are unique to TiO2 enrichment, and 3) those that are identified with both IMAC and TiO2 enrichment. We compared the numbers of singly and multiply phosphorylated peptides, the distribution of phospho-Serine (p-Ser), phospho-Threonine (p-Thr), and phospho-Tyrosine (p-Tyr) sites, and the detected charge states, the pI values and GRAVY values of the identified phosphopeptides for each subgroup. We also added two rounds of TiO2 after multi-step IMAC enrichment, or two rounds of IMAC after multi-step TiO2 enrichment to check whether the extra enrichment steps with different materials could recover more phosphopeptides and increase the overlap between IMAC and TiO2. The results showed that the additional rounds of TiO2 or IMAC were not enough to recover the phosphopeptides that were lost during the previous enrichment steps. This result suggests that the loss of phosphopeptides is due to irreversible binding to the IMAC or TiO2 beads.

MATERIALS AND METHODS

Cell Culture and Trypsin Digestion.

MCF-10A cells were cultured in DMEM/F12 (with 15 mM HEPES buffer) media containing 5% horse serum, 0.01 mg/mL insulin, 20 ng/mL EGF, 100 ng/mL cholera toxin, and 500 ng/mL hydrocortisone, as described previously 19. Cells were collected when they reached 80% confluence by rinsing twice with ice-cold PBS and lysing with 1mL/T175 flask of 8 M urea lysis buffer (containing 8 M urea, 75 mM NaCl, 50 mM Tris-HCl with pH8.2, 1 mM NaF, 1 mM-glycerophosphate, 1 mM sodium orthovanadate, 10 mM sodium pyro-phosphate, 1 mM PMSF, and 1 tablet of EDTA-free protease inhibitors cocktail for every 10 mL of lysis buffer). After three intervals of 1-minute sonication with a 2-minute rest in between, the mixture was centrifuged at 2500 g for 10 min and the supernatant was transferred to a new tube. Protein concentrations were measured with a bicinchoninic acid (BCA) assay kit (Thermo Scientific Pierce, Rockford, IL). Proteins were treated with 5 mM DTT for 25 min at 56 °C and then with 14 mM of IAA in the dark for 30 min at room temperature to allow alkylation of the cysteines. The reaction was stopped by adding another 5 mM DTT and reacting in the dark for 15 min. The protein mixture was then diluted with 25 mM Tris-HCl (pH 8.2) to achieve a final urea concentration of 1.8 M. Bovine trypsin (Sigma-Aldrich, St Louis, MO, USA) was added to protein mixture at a 50:1 protein to trypsin ratio (m/m) along with 1 mM CaCl2 and reacted overnight. After the trypsin digestion, unreacted trypsin was quenched with 0.4% (vol/vol) of TFA. Samples were then centrifuged at 2500 g for 10 min to remove any precipitates. The peptides were desalted with 500 mg reverse-phase tC18 Sep-Pak solid-phase extraction cartridges (Waters Corporation, Milford, MA) as reported previously 19.

IMAC Enrichment

The multi-step IMAC enrichment was performed as previously reported 19. Briefly, 30μL Phos-Select iron affinity gel (IMAC beads) (Sigma-Aldrich, St Louis, MO, USA) were washed 3 times with a 10× volume of IMAC binding buffer (40% ACN, 25 mM FA in H2O), and resuspended in IMAC binding buffer to make a 50% gel slurry. 20. Using the 100:1 (mg/μl) peptide-to-IMAC ratio optimized previously19, 3 mg of desalted, lyophilized whole-cell digest were resuspended in 360μL of IMAC binding buffer and incubated with 30 μl of bead slurry for 60min with vigorous shaking. After washing with 360 μl of binding buffer three times, phosphopeptides were eluted by shaking 5min in 40μL of 50 mM K2HPO4/NH4OH, pH10. The elution step was repeated three times to fully recover the phosphopeptides and the combined eluates were immediately acidified with 40μL of 10% formic acid. The second round of IMAC enrichment was performed with the supernatant of the first round of IMAC enrichment, while the third round of IMAC enrichment was performed with the supernatant from the second round of IMAC enrichment. At each step, the supernatant was applied directly to fresh beads as soon as it was removed from the previous aliquot of beads. The eluates from each round of the multi-step IMAC enrichment were then lyophilized and desalted with C18 ZipTips (Millipore, Billerica, MA) according to the manufacturer's instructions for LC-MS/MS analysis.

TiO2 Enrichment

TiO2 enrichment was performed as previously reported 21 with some modifications. The required amounts of TiO2 beads (GL Sciences, Tokyo, Japan) were washed with 500 μL loading buffer (65% ACN/2% TFA/saturated by glutamic acid) three times and resuspended in 200 μL loading buffer to make bead slurries. To optimize the TiO2 conditions, different amounts of TiO2 slurries were added to 3 mg of peptides according to the indicated peptide-to-TiO2 ratio (μg/ μg). The peptides were resuspended in 200 μL loading buffer, and incubated with the indicated amounts of TiO2 beads slurries at room temperature for 20 minutes while shaking. After incubation, the beads were then washed with 800 μL wash buffer I (65% ACN/0.5% TFA) once and 800 μL wash buffer II (65% ACN/0.1% TFA) twice. The TiO2 beads were then incubated with 200 μL elution buffer I (300 mM NH4OH/50% ACN) once and 200 μL elution buffer II (500 mM NH4OH/60% ACN) twice at room temperature for 20 minutes each time while shaking to elute the phosphopeptides. The eluates were combined and immediately acidified with formic acid. For multi-step TiO2 enrichment, the first round of TiO2 enrichment was completed with 3 mg of peptides from whole cell lysates and phosphopeptides were enriched with the optimized 1:2 peptide-to-TiO2 ratio (μg/ μg). The second enrichment was performed with supernatant from the first TiO2 enrichment, while the third round of enrichment was performed with the supernatant from the second TiO2 enrichment. The eluates from each round were then lyophilized and desalted with C18 ZipTips for LC-MS/MS analysis.

Mass Spectrometric Analysis

All samples were lyophilized and resuspended in 10 μl of MS loading buffer (1% HPLC grade ACN, 0.1% FA in HPLC grade water) and injected 2 μl for liquid chromatography electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS) analysis. A Q-Exactive mass spectrometer (Thermo Fisher Scientific, Bremen, Germany) coupled with a nanoACQUITY Ultra Performance LC (UPLC) system (Waters Corporation, Milford, MA) was used for all mass spectrometric analyses. Peptide separation was performed with a C18 reverse phase column (100 μm × 100 mm, 1.7 μm particle size, BEH130) (Waters Corporation, Milford, MA), with buffer A composed of 0.1% FA in water and buffer B containing 0.1% FA in ACN. A 73-min linear gradient from 3% to 30% buffer B was used for peptide separation, at a flow rate of 1200 nL/min. For the mass spectrometric analysis, the nanoelectrospray ion source was operated at a source voltage of 1.8 kV, with the ion transfer tube temperature set for 280 °C. The full MS scans were acquired in the Orbitrap mass analyzer with an m/z range of 350-1800, at the mass resolution of 70000 at m/z=200. Automatic gain control (AGC) target value was set to 1×106, with a maximum fill time of 250 ms. The top 12 most intense parent ions were selected for MS/MS with an isolation window of 2.0 m/z and fragmented under a normalized collision energy of 30%, with the AGC target value of 1×106 and the maximum fill time of 120 ms. The parent ions with unassigned charges or a charge state of z=1 were excluded from fragmentation. The intensity threshold for selection was set to 8.3×104. The fragmentation was performed in an HCD collision cell with a mass resolution of 35000 at m/z=200, and a dynamic exclusion period of 20 s. All samples were run in technical duplicate.

Data analysis

Database searching was performed with Proteome Discoverer 1.3 software (Thermo Fisher Scientific, Bremen, Germany) with MASCOT 2.2.4 against an in-house modified SwissProt human database by adding common contamination sequences (updated on 05/2012, 86758 sequences). Precursor peptide mass tolerance was set to 10 ppm and fragment ion mass tolerance was set to 0.02 Da. Carbamidomethylation of cysteine was a fixed modification, and the variable modifications were oxidation of methionine and phosphorylation of serine, threonine and tyrosine. Up to two missed cleavages were allowed for trypsin digestion and the peptide false discovery rate (FDR) was controlled at 0.01. Phosphopeptides with a PhosphoRS score higher than 0.99 were considered confidently localized 22. The circle Venn diagrams were generated by the online tool Venny (Oliveros, J.C. 2007, http://bioinfogp.cnb.csic.es/tools/venny/index.html). The three-circle proportional Venn diagrams were generated by CircleApplet, which can be downloaded from https://www.cs.kent.ac.uk/people/staff/pjr/EulerVennCircles/EulerVennApplet.html. The GRAVY values were calculated by the GRAVY calculator (http://www.gravy-calculator.de/). The pI values of phosphopeptides were calculated using pICalculator (https://trac.nbic.nl/picalculator/) 23 with ExPASy pKa values of amino acids. The pKa values for the phospho-groups were set as pKa1=2.12 and pKa2=7.21. The Motif analysis was performed with the online tool motif-x (http://motif-x.med.harvard.edu/) 24, 25, and motifs with the center of Ser were analyzed with a width of 15 amino acids and 20 occurrences, while motifs with the center of Thr and Tyr were analyzed with a width of 13 amino acids and 20 occurrences.

RESULTS AND DISCUSSION

Multi-step IMAC and TiO2 Enrich Similar Numbers of Phosphopeptides from Whole Cell Lysates

We have previously reported that IMAC enrichment performance depends significantly on the peptide-to-beads ratio 19. It was also reported in another publication that for TiO2 enrichment, the peptide-to-beads ratio is crucial for the phosphopeptide enrichment efficiency 21. In order to maximize the IMAC and TiO2 enrichment methods prior to comparing them, we first optimized the peptide-to-beads ratio for TiO2 as we did previously for IMAC. In our previous report, we showed that for IMAC enrichment, the 100:1 (μg:μl) peptide-to-beads ratio was the best ratio which allows IMAC to enrich phosphopeptides at the high phosphopeptide ratio of 62.3% 19. In this study, the results of optimization for the peptide-to-beads ratio for TiO2 show that TiO2 beads perform best with the peptide-to-beads ratio ranging from 1:2 to 1:8, which can achieve more than 80% phosphopeptide enrichment efficiency without losing total phosphopeptide identifications. An excess of solid phase did not improve phosphopeptide enrichment efficiency (Supporting Information Figure S1).

To fully extract phosphopeptides from the whole-cell lysate, we applied the multi-step enrichment strategy to TiO2 enrichment with the similar procedures for IMAC enrichment as described in our previous report 19. Briefly speaking, after the first round of enrichment with TiO2 (TiO2-1), a second round of enrichment (TiO2-2) was performed by enriching phosphopeptides from the supernatant of the first round of enrichment, and similarly a third round of enrichment (TiO2-3) was performed by enriching from the supernatant of the second round of enrichment (Figure 1A). In order to confirm that three rounds of enrichment was sufficient to enrich most of the phosphopeptides from the whole cell lysate, we evaluated the phosphopeptide number from each round of TiO2 enrichment, as well as their overlaps. TiO2-1 identified 4141 phosphopeptides. TiO2-2 enriched 1987 phosphopeptides, including 692 not found in TiO2-1. TiO2-3 added just 85 new identifications to the list out of the 564 phosphopeptides detected (Supporting Information Figure S2A). For the IMAC enrichment, 3724 phosphopeptides were found in I1, 1321 phosphopeptides were new in I2, and 230 phosphopeptides were unique to I3. As shown in Figure 1B, a total of 5272 unique phosphopeptides were identified with three rounds of IMAC enrichment, while a total of 4918 unique phosphopeptides were identified with three rounds of TiO2 enrichment, showing that both IMAC and TiO2 have comparable ability to enrich phosphopeptides from whole-cell lysates.

Figure 1.

Figure 1

Open in a new tab

Enrichment results from IMAC versus TiO2. (A) Workflow of multi-step IMAC and multi-step TiO2 enrichment. (B) Distribution of non-phosphopeptides and phosphopeptides identified with IMAC or TiO2 enrichment. (C) Overlaps of the non-redundant phosphopeptides identified with IMAC or TiO2 enrichment are shown in the Venn diagram. (D) Distribution of single- or multi-phosphopeptides in each subset. (E) Distribution of phosphopeptides with p-Ser, p-Thr, or p-Tyr sites that are identified in each subset.

Both enrichment techniques provided high specificity, but when considering the three rounds together, IMAC gave slightly better efficiency overall, with 72.3% phosphopeptides compared to 61.2% in TiO2 (Figure 1B). The first two IMAC enrichments were very efficient, with 79.4% phosphopeptides in I1 and 81.0% I2, with 51.4% in the third. TiO2 was highly selective in the first enrichment, at 86.1%, but yielded 49.1% and 16.7% phosphopeptides in the second and third rounds, respectively (Supporting Information Figure S2B). Even with the reduced specificity in the later rounds of enrichment, these steps still contributed 1462 phosphopeptides not found in either I1 or T1. All the above results indicate that three rounds of TiO2 or IMAC enrichment are sufficient to enrich most of the phosphopeptides from whole cell lysates.

Previous literature has reported that IMAC and TiO2 enrich distinct subsets of the phosphoproteome 16-18. In order to evaluate whether more complete enrichment could minimize the differences between IMAC and TiO2 enrichment, we next compared the differences of phosphopeptides that are enriched with the multiple rounds of TiO2 or IMAC. Comparing a single TiO2 and single IMAC enrichment shows 27% of the phosphopeptides were found by both techniques. Performing three rounds of enrichment increased this overlap only slightly; 34% of the phosphopeptides were identified with both IMAC and TiO2 enrichment (Figure 1C). Among the overall identified phosphopeptides, 2708 were unique to IMAC enrichment (I123-unique), 2354 were unique to TiO2 enrichment (T123-unique), and 2564 were identified with both methods (IT-Overlap).

IMAC-Unique, TiO2-Unique, and IMAC-TiO2-Overlap Subsets Shows Different Phospho-peptide Compositions

We further investigated the composition of the peptides unique to IMAC enrichment, unique TiO2 enrichment, and those found in the overlap subset. Although the total phosphopeptide number in each subset does not show much difference, there is a higher percentage of phosphopeptides with localized phosphosites in the IT-Overlap subset, indicating better spectral quality for phosphopeptides within the IT-Overlap subset.

Only those phosphopeptides with localization probability of ≥ 99% were considered for mono- or multi-phosphopeptide identification. The results show that IMAC enriched 3070 monophosphopeptides and 494 multi-phosphopeptides, while TiO2 enriched 3064 monophosphopeptides and 316 multi-phosphopeptides, indicating that IMAC is more efficient for enriching multi-phosphopeptides (Supplemental Information Figure S2C). In fact, the I123-Unique subset contains three times as many multi-phosphopeptides compared with the T123-Unique subset (Figure 1D), and the percentage of multi-phosphopeptides in I123-Unique subset is almost two times higher than that of T123-Unique or IT-Overlap subsets, with 18.4% compared to 7.4% and 10.6%, respectively. In comparison, the distribution of phospho-Ser, phospho-Thr, and phospho-Tyr does not show much difference among these three subsets (Figure 1E). The IT-Overlap subset shows a slightly higher composition of p-Ser phosphosites, which is most likely due to its higher amount of localized phosphosites (Figure 1E).

The peptide length, pI values, and GRAVY values of phosphopeptides within these three subsets were further analyzed. As shown in Figure 2, the I123-Unique subset shows higher affinity to longer phosphopeptides and it shows a higher amount of phosphopeptides longer than 20 amino acids compared to the T123-Unique and IT-Overlap subsets (Figure 2A). As for the pI values, the I123-Unique subset shows fewer phosphopeptides compared with the T123-Unique subset for phosphopeptides with pI values less than 4. In contrast, the I123-Unique subset shows more phosphopeptides compared with T123-Unique subset for phosphopeptides with pI values higher than 4 (Figure 2B). These results indicate that TiO2 has a higher affinity for acidic phosphopeptides compared with IMAC beads, which is consistent with existing literature 26, 27.

Figure 2.

Figure 2

Open in a new tab

Comparison of peptide length (A), pI values (B), and GRAVY values (C) for the I123-Unique, T123-Unique, and IT-Overlap subsets.

The hydrophobicity of phosphopeptides in each subset was determined by their GRAVY values. As shown in Figure 2C, the I123-Unique subset shows higher phosphopeptide amounts compared with the T123-Unique subset when GRAVY values are lower than −0.5. In contrast, the I123-Unique subset shows fewer phosphopeptides compared with the T123-Unique subset when GRAVY values are higher than −0.5 (Figure 2C). These results indicate that IMAC enrichment methods enrich more hydrophilic phosphopeptides compared with TiO2. Our findings are in contrast to a recent study that reported that enriching the IMAC supernatant and washes with TiO2 isolated longer and more multiply-phosphorylated peptides than IMAC enrichment alone 27. Given that our IMAC buffer contained a higher percentage of acetonitrile (40% for our buffer as compared to 0% in their loading buffer/ 25% in the wash), we speculate that the addition of the acetonitrile promotes binding of longer peptides to our IMAC beads in comparison to those used by Herring and coworkers.

In order to evaluate whether IMAC and TiO2 methods tend to enrich similar phosphopeptide structures, motif analysis was performed on the three subsets. Motifs with the center of Ser, Thr, and Tyr, were analyzed with Motif-X online tools 24,25 and summarized in Supporting Information Tables S1 and S2. Because the identified phosphopeptides with phosphorylated Tyr are relatively rare events, no specific motifs were identified that are centered at Tyr. The results show that the I123-Unique subset enriches substantially different phosphorylation motifs of phospho-Ser and phosphor-Thr compared with T123-Unique subset, and the differences are not fully covered by the IT-Overlap subset (Supporting Information Table S1, S2, & Figure S4).

Further Enrichment with IMAC from Multi-step TiO2 Enrichment, or Further Enrichment with TiO2 from Multi-step IMAC Enrichment Does Not Fully Recover Lost Phosphopeptides

We next evaluated whether the differences of phosphopeptides enriched by IMAC or TiO2 result from the enrichment step, or from the elution step. We used two rounds of TiO2 to further enrich phosphopeptides from the supernatant of three rounds of IMAC enrichment (I3T1, I3T2), and we used two rounds of IMAC to further enrich phosphopeptides from the supernatant of three rounds of TiO2 enrichment (T3I1, T3I2) (Figure 3A). The phosphopeptides identified by IMAC after TiO2 (I-T12, Figure 3B) and TiO2 after IMAC (T-I12, Figure 3C) do not show much overlap with those identified by three-step IMAC or three-step TiO2. They do overlap up to 50% with the IT-Overlap peptides, however. This indicates that the extra fourth and fifth rounds of enrichment primarily recover what remains of the most abundant peptides. These results suggest that the loss of phosphopeptide identifications is not due to incomplete enrichment from the cell lysate, but is more likely due to incomplete elution from IMAC or TiO2 materials.

Figure 3.

Figure 3

Open in a new tab

Distribution of phosphopeptides identified by two rounds of TiO2 following multi-step IMAC enrichment, or two rounds of IMAC following multi-step TiO2 enrichment. (A) Experimental workflow for multiple rounds of enrichment. (B) Overlap of TiO2 enriched phosphopeptides from IMAC supernatant with I123-Unique, T123-Unique, or IT-Overlap subsets. (C) Overlap of IMAC enriched phosphopeptides from TiO2 supernatant with I123-Unique, T123-Unique, or IT-Overlap subsets.

To evaluate the contribution of each round of enrichment to the total unique phosphopeptide identifications, for each round of enrichment, the total number of identified phosphopeptides and the number of unique phosphopeptides are summarized in Figure 4A. The results show that the first round of IMAC or TiO2 enrichment gave the highest number of both total phosphopeptide and unique phosphopeptide identifications, while the identification numbers decrease for the second and third round of enrichment with the same material. However, by switching the enrichment material after multi-step enrichment, the identification numbers increased again, although these additional enrichment steps contributed only a small percentage of the unique phosphopeptides. One unexpected result we observed pertains to the distribution of multi-phosphorylated peptides. In the first round of IMAC enrichment after multi-step TiO2 enrichment, the identified peptides were mostly multi-phosphorylated peptides (data not shown), and the total phosphopeptide numbers were relatively small (Figure 4A, T3I1). However, with a second round of IMAC treatment after multi-step TiO2 enrichment, the identified peptides were mostly single phosphopeptides (data not shown), significantly increasing the total phosphopeptide identification number (Figure 4A, T3I2). The number of identified peptides was comparable with the identification numbers for I3T1 and I3T2. The contribution of each round of enrichment for identification of unique phosphopeptides is summarized in Figure 4B. By integrating IMAC and TiO2 with the multi-step enrichment approach, a total of 8427 phosphopeptides were identified from ten fractions.

Figure 4.

Figure 4

Open in a new tab

Distribution of phosphopeptides identified in each round of TiO2 or IMAC enrichment. (A) Distribution of total phosphopeptides (blue) and unique phosphopeptides (pink) identified in each round of TiO2 or IMAC enrichment. (B) Contribution to overall identified unique phosphopeptides by each round of TiO2 or IMAC enrichment.

A recent report by Ruprecht and coworkers found that IMAC performed in an offline HPLC format identified more phosphopeptides than batch-mode TiO2 and tip-mode Ti-IMAC 8. The performance gap appeared to be primarily due to incomplete elution from beads used on the benchtop, which agrees with our observations in this study. Their results also demonstrated that a TiO2 –packed column retained phosphopeptides much more strongly than the IMAC column, required a more acidic loading buffer and more basic eluting buffer, and ultimately fractionated phosphopeptides over a longer window of elution time. Our finding that TiO2 enriches a much more acidic pool of peptides than IMAC is consistent with this observation. Finally, they concluded that the IMAC column and TiO2 chemistries are not complementary; complete fractionation of the IMAC-enriched pool yielded high overlap with the unfractionated TiO2 pool. For laboratories that do not have access to an extra HPLC or that are concerned about metal ion contamination in their instruments, batch-mode enrichments will continue to be popular low-cost, time-efficient alternatives. Our work provides a direct comparison of two popular batch-mode protocols and demonstrates that benchtop IMAC and TiO2 do not recover the same set of peptides. We recommend that laboratories using benchtop protocols consider using both chemistries in parallel with two to three rounds of sequential enrichment each. This approach should provide a good balance between recovery and instrument time, especially when further fractionation of each pool is planned.

CONCLUSION

In this study, we first optimized the peptide-to-beads ratio for the best performance of TiO2. Together with our previous findings studying IMAC, we conclude that three rounds of IMAC or TiO2 enrichment are sufficient to enrich most of the phosphopeptides from whole cell lysates. Comparison of multi-step IMAC and TiO2 enrichment shows that both enrichment methods have the equivalent ability to enrich phosphopeptides, with similar phosphopeptide identification numbers and similar enrichment efficiency. Further analysis of I123-Unique, T123-Unique, and IT-Overlap subsets shows that I123-Unique subset contains much higher amount of multi-phosphopeptides, while the composition of p-Ser, p-Thr, and p-Tyr shows no significant differences among these three subsets. The analysis of phosphopeptide length, pI values and GRAVY values shows that IMAC has higher affinity to longer phosphopeptides with the length of >20 amino acids, lower affinity to acidic phosphopeptides with pI values lower than 4, and higher affinity to hydrophilic phosphopeptides with GRAVY values lower than −0.5. The motif analysis also indicated that multi-step IMAC and TiO2 methods enrich different phosphorylation motifs. With further TiO2 enrichment from the supernatant of multi-step IMAC enrichment, or IMAC enrichment from the supernatant of multi-step TiO2 enrichment, the newly identified phosphopeptides did not increase the overlap between IMAC and TiO2 but instead mostly represented the most abundant species in the original lysate. From this we concluded that the IMAC and TiO2 appear to enrich distinct phosphopeptide subpopulations, not because of incomplete binding from the cell lysate, but more likely because of incomplete elution from the beads.

In conclusion, IMAC and TiO2 are relatively complementary methods for enriching phosphopeptides from whole-cell lysate. In our hands, the combination of IMAC and TiO2 enrichment identified a total number of 8427 phosphopeptides from whole cell lysate without any further fractionation. Based on these results, for a global phosphoproteome experiment where gaps in particular phosphopeptide pools must be avoided, a combination of the two enrichment chemistries with at least two rounds of enrichment would be expected to give the best coverage in the minimum amount of analysis time.

Supplementary Material

Supplemental Info

ACKNOWLEDGMENTS

The work was funded by Walther Cancer Foundation and the Notre Dame Harper Cancer Research Institute for postdoctoral funding for XY and salary support for ABH. This report was supported by the National Science Foundation (CAREER Award CHE-1351595) for ABH and the National Institutes of Health (1R01GM110406-01) for AS and ABH. The work was also funded by Department of Defense Visionary Postdoctoral Fellowship Award to XY (USAMRAA W81XWH-12-1-0412). We thank the staff from the Mass Spectrometry and Proteomics Facility at the University of Notre Dame, especially Dr. Bill Boggess, for their helpful discussions and advice, and for their help maintaining the mass spectrometry equipment. We also thank Dr. Susan Skube for her careful reading of the manuscript.

Footnotes

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

Supporting Information Available: Figures describe TiO2 peptide-to-bead optimization, TiO2 performance characteristics, and motif analyses. This information is available free of charge via the Internet at http://pubs.acs.org.

The authors declare no competing financial interest.

REFERENCES

  • 1.Engholm-Keller K, Larsen MR. Proteomics. 2013;13(6):910–931. doi: 10.1002/pmic.201200484. [DOI] [PubMed] [Google Scholar]
  • 2.Iliuk AB, Tao WA. Clin. Chim. Acta. 2013;420:23–27. doi: 10.1016/j.cca.2012.10.063. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Nilsson CL. Anal. Chem. 2012;84(2):735–746. doi: 10.1021/ac202877y. [DOI] [PubMed] [Google Scholar]
  • 4.Wang F, Song C, Cheng K, Jiang X, Ye M, Zou H. Anal. Chem. 2011;83(21):8078–8085. doi: 10.1021/ac201833j. [DOI] [PubMed] [Google Scholar]
  • 5.Fíla J, Honys D. Amino Acids. 2012;43(3):1025–1047. doi: 10.1007/s00726-011-1111-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Imamura H, Wakabayashi M, Ishihama Y. Semin. Cell. Dev. Biol. 2012;23(8):836–842. doi: 10.1016/j.semcdb.2012.05.007. [DOI] [PubMed] [Google Scholar]
  • 7.Andersson L, Porath J. Anal. Biochem. 1986;154(1):250–254. doi: 10.1016/0003-2697(86)90523-3. [DOI] [PubMed] [Google Scholar]
  • 8.Ruprecht B, Koch H, Medard G, Mundt M, Kuster B, Lemeer S. Mol. Cell. Proteomics. 2015;14(1):205–215. doi: 10.1074/mcp.M114.043109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Zhou H, Ye M, Dong J, Han G, Jiang X, Wu R, Zou H. J. Proteome Res. 2008;7(9):3957–3967. doi: 10.1021/pr800223m. [DOI] [PubMed] [Google Scholar]
  • 10.Ikeguchi Y, Nakamura H. Anal. Sciences. 1997;13(3):479–485. [Google Scholar]
  • 11.Matheron L, van den Toorn H, Heck AJ, Mohammed S. Anal. Chem. 2014;86(16):8312–8320. doi: 10.1021/ac501803z. [DOI] [PubMed] [Google Scholar]
  • 12.Zhang G, Neubert TA. J. Proteome Res. 2011;10(12):5454–5462. doi: 10.1021/pr200697x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Searleman AC, Iliuk AB, Collier TS, Chodosh LA, Tao WA, Bose R. Electrophoresis. 2014;35(24):3463–3469. doi: 10.1002/elps.201400022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Jayasundera KB, Iliuk AB, Nguyen A, Higgins R, Geahlen RL, Tao WA. Anal. Chem. 2014;86(13):6363–6371. doi: 10.1021/ac500599r. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Wang WH, Palumbo AM, Tan YJ, Reid GE, Tepe JJ, Bruening ML. J. Proteome Res. 2010;9(6):3005–3015. doi: 10.1021/pr901200m. [DOI] [PubMed] [Google Scholar]
  • 16.Cantin GT, Shock TR, Park SK, Madhani HD, Yates JR., 3rd Anal. Chem. 2007;79(12):4666–4673. doi: 10.1021/ac0618730. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Negroni L, Claverol S, Rosenbaum J, Chevet E, Bonneu M, Schmitter J MB. Analyt. Technol. Biomed. Life Sci. 2012:891–892. 109–112. doi: 10.1016/j.jchromb.2012.02.028. [DOI] [PubMed] [Google Scholar]
  • 18.Liang X, Fonnum G, Hajivandi M, Stene T, Kjus N H, Ragnhildstveit E, Amshey J W, Predki P, Pope R M. J. Am. Soc. Mass Spectrom. 2007;18(11):1932–1944. doi: 10.1016/j.jasms.2007.08.001. [DOI] [PubMed] [Google Scholar]
  • 19.Yue XS, Hummon AB. J. Proteome Res. 2013;12(9):4176–4186. doi: 10.1021/pr4005234. [DOI] [PubMed] [Google Scholar]
  • 20.Villen J, Gygi SP. Nat. Protoc. 2008;3(10):1630–1638. doi: 10.1038/nprot.2008.150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Li QR, Ning ZB, Tang JS, Nie S, Zeng R. J. Proteome Res. 2009;8(11):5375–5381. doi: 10.1021/pr900659n. [DOI] [PubMed] [Google Scholar]
  • 22.Taus T, Kocher T, Pichler P, Paschke C, Schmidt A, Henrich C, Mechtler K. J. Proteome Res. 2011;10(12):5354–5362. doi: 10.1021/pr200611n. [DOI] [PubMed] [Google Scholar]
  • 23.Gauci S, van Breukelen B, Lemeer SM, Krijgsveld J, Heck A. J. Proteomics. 2008;8(23-24):4898–4906. doi: 10.1002/pmic.200800295. [DOI] [PubMed] [Google Scholar]
  • 24.Chou MF, Schwartz D. Curr. Protoc. Bioinformatics. 2011 doi: 10.1002/0471250953.bi1315s35. Chapter 13:Unit 13,15-24. [DOI] [PubMed] [Google Scholar]
  • 25.Schwartz D, Gygi SP. Nature Biotechnology. 2005;23(11):1391–1398. doi: 10.1038/nbt1146. [DOI] [PubMed] [Google Scholar]
  • 26.Thingholm TE, Jensen ON, Robinson PJ, Larsen MR. Mol. Cell. Proteomics. 2008;7(4):661–671. doi: 10.1074/mcp.M700362-MCP200. [DOI] [PubMed] [Google Scholar]
  • 27.Herring LE, Grant KG, Blackburn K, Haugh JM, Goshe MB. J. Chromatogr. B. 2015;988:166–174. doi: 10.1016/j.jchromb.2015.02.017. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Info
NIHMS760470-supplement-Supplemental_Info.docx (3.1MB, docx)

RESOURCES