Rapid Combinatorial ERLIC−SCX Solid-Phase Extraction for In-Depth

Rapid Combinatorial ERLIC−SCX Solid-Phase Extraction for In-DepthPhosphoproteome AnalysisMostafa Zarei,*,†,‡,⊥ Adrian Sprenger,†,‡,§,⊥ Christine Gretzmeier,†,‡ and Joern Dengjel*,†,‡,§,∥

†Freiburg Institute for Advanced Studies (FRIAS), School of Life Sciences-LifeNet, University of Freiburg, Albertstr. 19, 79104Freiburg, Germany‡ZBSA Center for Biological Systems Analysis, University of Freiburg, Habsburgerstr. 49, 79104 Freiburg, Germany§Department of Dermatology, University Freiburg Medical Center, Hauptstr. 7, 79104 Freiburg, Germany∥BIOSS Centre for Biological Signalling Studies, University of Freiburg, Schanzlestr. 18, 79104 Freiburg, Germany

*S Supporting Information

ABSTRACT: Protein phosphorylation is an important mechanismof cellular signaling, and many proteins are precisely regulatedthrough the interplay of stimulatory and inhibitory phosphorylationsites. Phosphoproteomics offers great opportunities to unravel thiscomplex interplay, generating a mechanistic understanding of vitalcellular processes. However, protein phosphorylation is substoichio-metric and, in particular, peptides carrying multiple phosphorylationsites are extremely difficult to detect in a highly complex mixture ofabundant nonphosphorylated peptides. Chromatographic methodsare employed to reduce sample complexity and thereby significantlyincrease the number of phosphopeptide identifications. Wepreviously demonstrated that combinatorial strong cation ex-change−electrostatic repulsion−hydrophilic interaction chromatog-raphy yields a surplus in overall identifications of phosphopeptides compared with single chromatographic approaches. Here wepresent a simple and rapid strategy implemented as solid-phase extraction not requiring specific instrumentation such as off-lineHPLC systems. It is inexpensive, adaptable for high and low amounts of starting material, and saves time by allowing multiplexedsample preparation without any carry-over problem.

KEYWORDS: SPE, ERLIC, SCX, phosphoproteomics, TiO2, chromatography, mass spectrometry

■ INTRODUCTION

Protein phosphorylation represents one of the most importantmechanisms of cellular signal transduction,1 influencing amongothers protein stability, activity, subcellular localization, andaffinity.2 Phosphorylation introduces a polar charge that caninduce a conformational change in proteins or can block or freebinding sites. Frequently, proteins are regulated through theconcerted action of several phosphorylation sites.3 Massspectrometry (MS) enables unbiased large-scale analyses ofthe proteome and its post-translational modification.4 Prior toMS analysis, proteins are commonly digested, as peptides canbe more easily separated and ionized. Phosphorylation sites arerecognized through a mass shift of 79.9963 Da5 and can bespecifically addressed in mass spectrometric analysis strategies,such as multistage activation (MSA),6 higher energy collisionaldissociation (HCD), or collision-induced dissociation (CID).7

If the MS/MS spectrum is of certain quality, a uniquelocalization of the phosphorylation site within a peptide ispossible.8

However, in-depth proteome analysis is hampered by theenormous complexity of whole cell lysate digests and the highdynamic range of protein abundances. The analysis of the

phosphoproteome adds additional challenges. Because potentialphosphorylation sites are generally occupied at substoichio-metric levels, phosphopeptides are low abundant and in acomplex mixture of excess nonphosphorylated peptides. Onlywith phospho-affinity purification methods, such as immobi-lized metal ion chromatography (IMAC)9 and TiO2 affinitychromatography,10,11 phosphopeptides can be enriched to anextent that enables efficient MS-based phosphoproteomeanalysis. 30% of all proteins are believed to carry at least onephosphorylation site.4,12 The many possible occupationcombinations of peptides with multiple phosphylation sitestremendously increase the complexity of the sample. Rapidtechnological advances in mass spectrometers account for thisby increasing scanning speed and sensitivity.13−15 However,biological sample complexity still exceeds the separation powerand acquisition speed of state-of-the-art LC−MS combina-tions.16 Thus, for in-depth analyses, fractionation and enrich-ment of phosphopeptides prior to LC−MS/MS analysis is stillstate-of-the-art.17,18 Different physicochemical traits of phos-

Received: August 1, 2013Published: October 22, 2013

Technical Note

pubs.acs.org/jpr

© 2013 American Chemical Society 5989 dx.doi.org/10.1021/pr4007969 | J. Proteome Res. 2013, 12, 5989−5995

Dow

nloa

ded

by N

AT

L L

BR

Y O

F SE

RB

IA o

n Se

ptem

ber

11, 2

015

| http

://pu

bs.a

cs.o

rg

Pub

licat

ion

Dat

e (W

eb):

Nov

embe

r 4,

201

3 | d

oi: 1

0.10

21/p

r400

7969

pubs.acs.org/jpr

phopeptides can be addressed for optimal fractionation, forexample, hydrophobicity, hydrophilicity, or charge.17,19 Thecombinatorial use of two chromatographic methods addressinginverse or complementary physicochemical characteristics ofphosphopeptides consequently yields an improved fractiona-tion. Offline SCX fractionation coupled to online RP−RP(high- and low-pH), RP−TiO2−RP, or SCX−WAX all havebeen described in recent studies.20−23

We recently proposed the combinatorial use of ERLIC andSCX to achieve efficient retention and fractionation, especiallyof very low abundant multiple phosphorylated peptides.24 Bothmethods are performed at low pH (<2.7) to uncharge acidicpeptides and retain the negative charges of phosphate groups.Under these conditions, most tryptic peptides have a solutioncharge state of +2, decreasing by −1 for each attachedphosphate group.12,25 Hence, the positively charged stationaryphase of ERLIC favors retention and separation of negativelycharged multiphosphorylated peptides, while the negativelycharged stationary phase of SCX favors retention andseparation of unphosphorylated or singly phosphorylatedpeptides.26,27 In this way, the very low abundant multiplephosphorylated peptides are separated from the more abundantsingly phosphorylated peptides, increasing their identificationrate immensely.A routine application of this method in average life science

laboratories is hampered by several factors: (1) Expertise in theuse of FPLC systems in different chromatographic modes isrequired. (2) Switching between different chromatographicmodes is laborious and time-consuming;. (3) The use of a

FPLC generally requires large sample amounts. To addressthese drawbacks, we transfer the principle of combinatorialphosphopeptide fractionation by ERLIC−SCX to a self-packedsolid-phase extraction (SPE) setup and demonstrate highphosphopeptide separation and identification rates for single aswell as multiple phosphorylated peptides. The SPE procedure isinexpensive and easy-to-use, enabling a quick implementationin any laboratory.28 Moreover, it allows parallel handling ofERLIC and SCX samples saving a substantial amount of timecompared with traditional FPLC applications.

■ MATERIAL AND METHODS

Chemicals

All solvents, acetonitrile (ACN) (Wako, Neuss, Germany),acetic acid (LGC Promochem, Wesel, Germany), formic acid(FA), and trifluoroacetic acid (TFA) (Merck, Darmstadt,Germany), were LC-MS grade. Sodium deoxycholate, dithio-threitol (DTT), ammonium bicarbonate (ABC), iodoacetamide(IAA), lactic acid, triethylamine (TEA), sodium dihydrogenphosphate (NaH2PO4), and sodium acetate were all purchasedfrom Sigma Aldrich (Munich, Germany). Potassium dihydro-gen phosphate (KH2PO4) and potassium chloride (KCl) werepurchased from Merck. Methylphosphonic acid (MePO4) wasfrom Alfa Aesar (Karlsruhe, Germany). Sequencing gradetrypsin was purchased from Promega (Mannheim, Germany).Water used in the experiments was prepared using a Milli-Qsystem (Millipore, Bedford, MA).

Figure 1. Experimental workflow. Proteins in whole cell lysate were digested using trypsin. Peptides were fractionated by ERLIC−SPE and SCX−SPE, respectively. The FT from the ERLIC column was loaded onto the SCX−SPE column. In total, five fractions were collected from each step,plus the FT from SCX. All 10 fractions were subjected to one-time TiO2-based phosphopeptide enrichment, the FT to three consecutive times,yielding 13 phosphopeptide samples that were analyzed by RP−LC−MS/MS.

Journal of Proteome Research Technical Note

dx.doi.org/10.1021/pr4007969 | J. Proteome Res. 2013, 12, 5989−59955990

Dow

nloa

ded

by N

AT

L L

BR

Y O

F SE

RB

IA o

n Se

ptem

ber

11, 2

015

| http

://pu

bs.a

cs.o

rg

Pub

licat

ion

Dat

e (W

eb):

Nov

embe

r 4,

201

3 | d

oi: 1

0.10

21/p

r400

7969

Sample Preparation

HeLa cells were grown to near confluence in 15 cm dishes andscraped and lysed in 1% sodium deoxycholate in 50 mM ABCbuffer, and the lysate was clarified by centrifugation.29,30

Supernatants were reduced with 1 mM DTT and alkylatedwith 5.5 mM IAA for 45 min at room temperature in darknessbefore being digested with trypsin (1:75 trypsin/protein)overnight at 37 °C. Digests were acidified to 1% FA, andundigested proteins and deoxycholate were removed bycentrifugation. The protein concentration was measured usingthe bicinchoninic acid assay (Thermo Fisher Scientific, Bonn,Germany), and tryptic digests were stored at −80 °C for furtheruse.

Preparation of ERLIC− and SCX−SPE Material

For fractionation by ERLIC− and SCX−SPE, respective bulkmaterial was self-packed into 1 mL pipet tips with punchedWhatman paper as frit in the front part (Figure 1). One g ofsolid-phase slurry of each ERLIC Poly-WAX LP (5 μm particlesize, 300 Å pore size; PolyLC, Columbia, USA) and SCXPolySULFOETHYL A (5 μm particle size, 300 Å pore size;PolyLC), was prepared according to the following protocol: (1)Transfer both solid-phase materials to separate 50 mL reactiontubes and wet beads using 25 mL of methanol by soft rotationat room temperature (RT) for 10 min. (2) Centrifuge tubes at0.2 rcf and wash beads with 5 mL of water. (3) Wet beads with40 mL of 0.2 M sodium dihydrogen phosphate (NaH2PO4) and0.3 M sodium acetate in water for 1 h at RT by soft mixing. (4)Wash beads with 5 mL of water and add water to the solidphase to reach the final volume of 4 mL. For each SPEexperiment, we used 1 mL of these slurries that arerepresentative of 250 mg of respective starting materials. (5)Transfer 1 mL of solid phase slurry to the blocked pipet tipsand flush it with 1 mL of respective solvent B and (6) by aminimum of 5 mL of respective solvent A. (7) Load the sampleto the beads and elute with different ratios of solvent B/A (seeworkflow section). In all steps, the liquid was eluted bycentrifugation at 0.8 rcf. Two cm of the top part of pipet tipswas cut to fit them into a tabletop centrifuge.

Workflow of ERLIC−SCX−SPESix mg of HeLa cell tryptic protein digest was acidified by FA toa pH under 3, diluted to 1 mL with solvent A of ERLIC, andloaded a minimum of two times to an equilibrated ERLIC−SPEtip. The flow through (FT) was collected, concentrated toreduce the ACN concentration to 30%, acidified with FA, andloaded a minimum of two times to an equilibrated SCX−SPEcolumn. Afterward, peptides on both SPE columns werefractionated in five steps by 1 mL of solvent B/solvent A indifferent ratios. For ERLIC, 20, 40, 60, 80, and 100% of solventB were chosen, while for SCX, 10, 20, 30, 40 and 100% solventB provided greater fractionation efficiency. These percentageswere determined in setup experiments (data not shown).For ERLIC, 10 mM sodium methylphosphonate (Na-

MePO4) with 70% ACN, pH 2.0, and 200 mM triethylaminephosphate (TEAP) with 60% ACN, pH 2.0 were used assolvent A and solvent B, respectively. In SCX, 30% ACNcontaining 5 mM KH2PO4, pH 2.7, was used as solvent A and30% ACN containing 5 mM KH2PO4 and 300 mM KCl, pH2.7, was used as solvent B.

TiO2 Enrichment

Ten μL of a 30% (V/V) TiO2 slurry (MZ-Analysentechnik,Mainz, Germany) in 30 mg/mL lactic acid was added to each

fraction and FT and incubated for 30 min at roomtemperature.31 Beads were washed with 100 μL of 10% ACNand 1% TFA, followed by 100 μL of 80% ACN and 1% TFAand finally 100 μL of water. Phosphopeptides were eluted using25% ammonium hydroxide in 20 and 40% ACN, respectively.Eluted phosphopeptides were dried to <5 μL and resuspendedin 15 μL of 0.5% acetic acid in water for analysis. Five μL ofeach fraction was used for MS analysis. For consecutiveincubations, the peptide-beads slurry was spun down, and thesupernatant was incubated with another aliquot of freshlyprepared TiO2 beads. Average enrichment of TiO2 on differentchromatography systems was calculated as the number ofdetected unique phosphopeptides divided by the total numberof detected unique peptides per fraction.

Mass Spectrometric Analysis

Mass spectrometric measurements were performed on an LTQOrbitrap XL mass spectrometer (Thermo Fisher Scientific,Bremen, Germany) coupled to an Agilent 1200 (AgilentTechnologies, Waldbronn, Germany) or an Eksigent 2Dnanoflow-HPLC (AB Sciex, Darmstadt, Germany). HPLC-column tips (fused silica) with 75 μm inner diameter (NewObjective, Woburn, MA) were self-packed with Reprosil-Pur120 ODS-3 (Dr. Maisch, Ammerbuch, Germany) to a length of20 cm. Samples were applied directly onto the column withoutprecolumn. A gradient of A [0.5% acetic acid in water] and B[0.5% acetic acid in 80% ACN/water] with increasing organicproportion was used for peptide separation (loading of samplewith 2% of solvent B; first separation ramp from 2 to 35% Bwithin 100 min, and second ramp from 35 to 80% B within 20min). The flow rate was 250 nL/min and for sample application500 nL/min. Two technical replicates with identical settingswere performed. The mass spectrometer was operated in thedata-dependent mode and switched automatically between MS(max. of 1 × 106 ions) and MS/MS. Each MS scan wasfollowed by a maximum of five MS/MS scans in the linear iontrap using 35% collision energy and a target value of 5000.Parent ions with a charge state of z = 1 and unassigned chargestates were excluded for fragmentation. Further MS/MSsettings were: repeat duration (30 s), repeat count (1),exclusion duration (45 s), and isolation width (2). For MS/MS, wideband activation and multi stage activation wereenabled with the neutral loss mass list of singly, doubly, andtriply phosphorylated peptides. Remaining settings were set todefault. The mass range for MS was m/z = 350 to 2000, andsignal threshold was 1000. The resolution was set to 60 000.Mass-spectrometric parameters were as follows: spray voltage2.3 kV; no sheath and auxiliary gas flow; ion-transfer tubetemperature 200 °C.Data Analysis

All raw files were analyzed with the software MaxQuant(version 1.0.1.18).32,33 Cysteine carbamidomethylation wasselected as fixed modification; methionine oxidation, proteinN-terminal acetylation, and phosphorylation on serine,threonine, and tyrosine were selected as variable modifications.Up to three miscleavages were allowed. Precursor ion masstolerance was 6 ppm, and fragment ion mass tolerance was 0.5Da for MS/MS spectra. A false discovery rate (FDR) of 1% anda minimum length of six amino acids were used forphosphopeptide and site identifications. The “Evidence” tablewas basis for peptide analyses. Two peptides identified by twosequence spectra were regarded as unique if they (a) differed intheir amino acid sequence or (b) differed in their



Dow

nloa

ded

by N

AT

L L

BR

Y O

F SE

RB

IA o

n Se

ptem

ber

11, 2

015

| http

://pu

bs.a

cs.o

rg

Pub

licat

ion

Dat

e (W

eb):

Nov

embe

r 4,

201

3 | d

oi: 1

0.10

21/p

r400

7969

phosphorylation site(s). Different charge states were not takeninto account.

■ RESULTS AND DISCUSSIONFor the combinatorial ERLIC−SCX−SPE workflow forphosphopeptide fractionation, HeLa cell lysate was used(Figure 1). Six mg cells were lysed using sodium deoxycholate,resulting in 1 mL of cell lysate and digested with trypsin.Peptides were taken up in ERLIC buffer A and loaded onto theERLIC−SPE column. (See the Material and Methods fordetails.) The ERLIC FT was concentrated in a SpeedVac toreduce the concentration of acetonitrile to 30%, acidified by FA,loaded onto the SCX−SPE column, and the SCX FT was againcollected. In contrast with an HPLC-based workflow, it shouldbe considered that in the SPE approach the columns are packedonly by centrifugation and not by high pressure. The resultingdead volume between the beads is rather high and results in lessefficient peptide binding. Therefore, the samples were reloadedup to three times onto the same tip to improve binding.Peptides from ERLIC−SPE and SCX−SPE columns were eacheluted in five steps using increasing percentages of respectiveelution buffers. (See the Material and Methods for details.)Each fraction was subjected to one TiO2-based phosphopeptideenrichment step, except the SCX FT, which was incubatedthree times. The resulting 13 phosphopeptide samples wereanalyzed by LC−MS/MS using collision-induced dissociationand multistage activation as fragmentation technique. Thirteen145 min runs gave rise to a total MS measurement time of ∼32h per experiment. MaxQuant and Andromeda were employedfor peptide identification.33 All experiments were performedminimally in two technical replicates, starting from theERLIC−SPE step. (See Supplementary Figure S1 and TableS1 in the Supporting Information.)In the combinatorial workflow, 13 fractions yielded a total of

51 244 peptide−spectrum matches (PSMs), of which 42 012could be matched to phosphopeptides. 9952 of thesephosphopeptides were unique, and of these 2137 were multiplyphosphorylated (≥3p). The peptides gave rise to more than5670 phosphorylation sites on 2198 proteins. The complete listof identified unique phosphopeptides is available as Supple-mentary Table S2 in the Supporting Information. Phosphopep-tide enrichment efficiency was on average 93% for the ERLICfractions and 71% for the SCX fractions. When measurementtime is scarce, SCX fractions 4 and 5 can be omitted with only7% data loss (Figure 2A). We recommend not to omit anyfractions of the ERLIC approach because more than 80% of theunique phosphopeptides found in fractions 4 and 5 are usuallyoverlooked, as they are low-abundant, carrying multiplephosphate groups.To confirm separation according to the inherent affinity of

negative charges in ERLIC and positive charges in SCX,respectively, we analyzed the average solution net charge ofidentified peptides in each fraction. In the ERLIC fractions, weobserve the expected decrease in average peptide net chargewith increasing fraction number; in the SCX fractions, weobserve the expected increase (Figure 2B). An increase incharge state in the consecutive FT phosphopeptide enrichmentsteps can also be observed, as multiphosphorylated peptideshave a stronger affinity to TiO2 and are consequently depletedfirst (data not shown). Of all phosphopeptides identified in allERLIC and SCX fractions, 69% are found in just one fraction,and 20% are shared between two fractions, indicating a fairfractionation quality throughout the two different chromato-

graphic methods (Figure 2C). To get a more detailed overviewof fraction overlap between both chromatographic methods, wevisualize the overlap of each fraction with all other fractions.For this purpose, we assigned phosphopeptides primarily tothat fraction in which they were found to be most abundantbased on median-normalized intensities (Figure 2D). We foundthe largest overlap between the FT and the first fractions ofERLIC and SCX, respectively. This can be expected becausethese fractions select for similar solution charge states (Figure2B). Also, peptides eluting in the very first fraction of anychromatographic separation generally have the lowest bindingaffinity to the solid phase and consequently can also be foundto a large portion in the respective FT.ERLIC and SCX are complementary, as indicated by the

small overlap of unique phosphopeptides between the twoconsecutive separation steps. 3757 (38%) unique phosphopep-tides were found in the ERLIC fractions, 4653 (47%) werefound in the SCX fractions, and only 1542 (15%) were found infractions of both methods (Supplementary Table S2 in theSupporting Information) (Figure 3A). The orthogonality of thetwo approaches is also highlighted by their ability to separatemultiply and singly phosphorylated peptides, respectively(Figure 3B). In ERLIC, the fraction size of multiphosphory-lated peptides (≥3p) is on average 31% and goes up to 92% infraction five. In contrast with that, SCX retains maximally 11%multiphosphorylated peptides. ERLIC fractions 2−4 are themost interesting in this regard, contributing each on average336 unique multiphosphorylated peptides to the whole data set.Fraction 5 consists of 14% of peptides carrying more than four

Figure 2. Yield of the combinatorial ERLIC−SCX−SPE workflow.(A) Number of unique phosphopeptides (green bars), uniqueunphosphorylated peptides (gray bars) in each fraction, andcumulative count (red line) are shown. pp = phosphopeptide. (B)Average solution net charge state of peptides in respective fractions,(C) overall peptide overlap between fractions as the percentage ofpeptides that are found in 1, 2, 3, or more fractions, and (D) detailedoverview of peptide overlap, showing how peptides assigned to aspecific fraction (by median-normalized intensity) distribute across allother fractions.



Dow

nloa

ded

by N

AT

L L

BR

Y O

F SE

RB

IA o

n Se

ptem

ber

11, 2

015

| http

://pu

bs.a

cs.o

rg

Pub

licat

ion

Dat

e (W

eb):

Nov

embe

r 4,

201

3 | d

oi: 1

0.10

21/p

r400

7969

phosphorylation sites (Figure 3C). ERLIC fractions containedin total 1685 (32%) unique multiphosphorylated peptides,while SCX fractions contained only 576 (9%) uniquemultiphosphorylated peptides. The combined ERLIC−SCXfractionation yields a total of 2137 (22%) unique multi-phosphorylated peptides (Figure 3D). Of these, 79% arecontributed by the ERLIC part and only 21% by the SCX part.Of all singly/doubly phosphorylated unique peptides, themajority is identified in the SCX fractions (60%; Figure 3E).To compare the combinatorial fractionation using SPE with a

respective FPLC-based approach, we used data from thecurrent study and from our already published work.24 In bothcases, we used the same amount of sample and the sameinstrumental setup. Both approaches yielded very similar resultswith an identification of 386 phosphopeptides per hour in caseof the FPLC approach versus 377 phosphopeptides per hour inthe case of the SPE approach. The differences in total numbersof identified phosphopeptides (6433 versus 9552) can beexplained by the different LC gradient lengths. However, with22% unique multiple phosphorylated peptides, we achieved athree-fold increase in the SPE workflow compared with theFPLC workflow. One reason for this might be the five timessmaller amount of stationary phase used in the SPE workflow.The sample amount of 6 mg likely exhausted the bindingcapacity of the ERLIC stationary phase primarily capturing thehighly affine multiphosphorylated peptides. Lower abundanceof singly phosphorylated peptides in the respective fractionsincreases MS fragmentation frequency of multiphosphorylatedpeptides, and hence identification rates. This demonstrates thatadaptation of the amount of stationary phase to the amount ofsample is an important factor to fine-tune the complementaryuse of ERLIC−SCX−SPE. The MaxQuant analysis of theERLIC fractions identified with 30% peptide-spectrum matchesrelatively few peptides from the recorded MS/MS spectra. This

suggests that the choice of the MS fragmentation technique isanother opportunity for fine-tuning. Although more than 31%of the detected phosphopeptides in the ERLIC fractions weremultiphosphorylated, we think this number could be evenfurther increased by use of ETD or HCD as a complementaryfragmentation technique to CID.34,35

A major advantage of the SPE-based workflow is itssuitability for large as well as small amounts of startingmaterial. In contrast with the tubing in a FPLC-based workflow,the SPE setup presents only little hydrophobic surfaces thatpotentially adsorb peptides, and hence it requires less startingmaterial. To demonstrate this, we performed a 1D SPE usingSCX only. We loaded 500 μg of HeLa cell lysate onto 200 mgof the respective stationary phase packed in a 1 mL pipet tip.Five fractions plus FT were collected, phosphopeptides wereenriched using TiO2, and samples were analyzed by LC−MS/MS. We identified 1044 unique phosphopeptides (Supple-mentary Table S3 in the Supporting Information), which is afair amount in our instrumental setup.

■ CONCLUSIONSConsidering the importance of protein phosphorylation inphysiological and pathological signal transduction, phospho-proteome analysis holds great value in understanding theunderlying molecular mechanisms leading cellular phenotypes.The combinatorial ERLIC−SCX−SPE setup described heremakes sample preparation for in-depth phosphoproteomeanalysis accessible for a broad range of biological laboratorieswithout the need for detailed expertise in chromatographicmethods and access to respective equipment. It is inexpensive,suitable for limited samples amount, for example, from primarycells, and saves time and labor by allowing multiplexed samplepreparation. With 22% of all identified phosphopeptides beingmultiple phosphorylated (≥3p), the method enables the large-

Figure 3. Complementarity of ERLIC and SCX: (A) Overlap of unique phosphopeptides between ERLIC and SCX. (B) Percentage of nonuniquepeptides with different degree of phosphorylation in each single fraction. (C) Number of unique multiphosphorylated peptides identified in differentERLIC fractions plus cumulative count (pp = phosphopeptide). (D) Percentage of unique peptides with different degree of phosphorylation inERLIC fractions, SCX fractions + FT, and in the combined workflow. (E) Pie chart illustrating the contribution of ERLIC and SCX to theidentification of singly/doubly and multiple phosphorylated unique peptides.



Dow

nloa

ded

by N

AT

L L

BR

Y O

F SE

RB

IA o

n Se

ptem

ber

11, 2

015

| http

://pu

bs.a

cs.o

rg

Pub

licat

ion

Dat

e (W

eb):

Nov

embe

r 4,

201

3 | d

oi: 1

0.10

21/p

r400

7969

scale study of protein regulation controlled by the interplay ofseveral phosphorylation sites.

■ ASSOCIATED CONTENT

*S Supporting Information

In addition to supporting Figure S1, three supporting tableswith identified phosphopeptides (Supporting Table S1[replicate experiment], Supporting Table S2 [ERLIC−SCX−SPE], and Supporting Table S3 [down scaling of SCX−SPE])are submitted with this manuscript. This material is availablefree of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Authors

*J.D.: Tel: +49-761-203-97208. Fax: +49-761-203-8456. E-mail: [email protected].*M.Z.: Tel: +49-761-203-97220. E-mail: [email protected].

Author Contributions⊥M.Z. and A.S. contributed equally to this work.

Notes

The authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThe research leading to these results has received funding fromthe Excellence Initiative of the German Federal and StateGovernments through FRIAS and BIOSS, from the FederalMinistry of Education and Research through GerontoSys II_NephAge (031 5896 A), and from the German ResearchFoundation, DFG (DE 1757/2-1). We thank Eksigent/ABSciex for their generous gift of a 2D nanoLC and Remco vanSoest for his technical advice for maintenance and troubleshooting.

■ REFERENCES(1) Hunter, T. The age of crosstalk: phosphorylation, ubiquitination,and beyond. Mol. Cell 2007, 28, 730−738.(2) Johnson, L. N.; Barford, D. The effects of phosphorylation on thestructure and function of proteins. Annu. Rev. Biophys. Biomol. Struct.1993, 22, 199−232.(3) Halbach, S.; et al. Alterations of Gab2 signalling complexes inimatinib and dasatinib treated chronic myeloid leukaemia cells. CellCommun. Signaling 2013, 11, 30.(4) Olsen, J. V.; et al. Global, in vivo, and site-specificphosphorylation dynamics in signaling networks. Cell 2006, 127,635−648.(5) Boersema, P. J.; Mohammed, S.; Heck, A. J. R. Phosphopeptidefragmentation and analysis by mass spectrometry. J Mass Spectrom2009, 44, 861−878.(6) Schroeder, M. J.; Shabanowitz, J.; Schwartz, J. C.; Hunt, D. F.;Coon, J. J. A neutral loss activation method for improvedphosphopeptide sequence analysis by quadrupole ion trap massspectrometry. Anal. Chem. 2004, 76, 3590−3598.(7) Jedrychowski, M. P.; et al. Evaluation of HCD- and CID-typefragmentation within their respective detection platforms for murinephosphoproteomics. Mol. Cell Proteomics 2011, 10, M111.009910.(8) Marx, H.; et al. A large synthetic peptide and phosphopeptidereference library for mass spectrometry-based proteomics. Nat.Biotechnol. 2013, 31, 557−564.(9) Andersson, L.; Porath, J. Isolation of phosphoproteins byimmobilized metal (Fe3+) affinity chromatography. Anal. Biochem.1986, 154, 250−254.

(10) Kuroda, I.; Shintani, Y.; Motokawa, M.; Abe, S.; Furuno, M.Phosphopeptide-selective column-switching RP-HPLC with a titaniaprecolumn. Anal. Sci. 2004, 20, 1313−1319.(11) Larsen, M. R.; Thingholm, T. E.; Jensen, O. N.; Roepstorff, P.;Jørgensen, T. J. D. Highly selective enrichment of phosphorylatedpeptides from peptide mixtures using titanium dioxide microcolumns.Mol. Cell Proteomics 2005, 4, 873−886.(12) Beausoleil, S. A.; et al. Large-scale characterization of HeLa cellnuclear phosphoproteins. Proc. Natl. Acad. Sci. U.S.A. 2004, 101,12130−12135.(13) Kocher, T.; Swart, R.; Mechtler, K. Ultra-high-pressure RPLChyphenated to an LTQ-Orbitrap Velos reveals a linear relationbetween peak capacity and number of identified peptides. Anal. Chem.2011, 83, 2699−2704.(14) Michalski, A.; et al. Mass spectrometry-based proteomics usingQ Exactive, a high-performance benchtop quadrupole Orbitrap massspectrometer. Mol. Cell Proteomics 2011, 10, M111.011015.(15) Oppermann, F. S.; et al. Comparison of SILAC and mTRAQquantification for phosphoproteomics on a quadrupole orbitrap massspectrometer. J. Proteome Res. 2013, 12, 4089−4100.(16) Michalski, A.; Cox, J.; Mann, M. More than 100,000 detectablepeptide species elute in single shotgun proteomics runs but themajority is inaccessible to data-dependent LC-MS/MS. J. Proteome Res.2011, 10, 1785−1793.(17) Engholm-Keller, K.; Larsen, M. R. Technologies and challengesin large-scale phosphoproteomics. Proteomics 2013, 13, 910−931.(18) Rigbolt, K. T. G.; Blagoev, B. Quantitative phosphoproteomicsto characterize signaling networks. Semin. Cell Dev. Biol. 2012, 23,863−871.(19) McNulty, D. E.; Annan, R. S. Hydrophilic interactionchromatography reduces the complexity of the phosphoproteomeand improves global phosphopeptide isolation and detection. Mol. CellProteomics 2008, 7, 971−980.(20) Hennrich, M. L.; Groenewold, V.; Kops, G. J. P. L.; Heck, A. J.R.; Mohammed, S. Improving depth in phosphoproteomics by using astrong cation exchange-weak anion exchange-reversed phase multi-dimensional separation approach. Anal. Chem. 2011, 83, 7137−7143.(21) Lemeer, S.; et al. Online automated in vivo zebrafishphosphoproteomics: from large-scale analysis down to a singleembryo. J. Proteome Res. 2008, 7, 1555−1564.(22) Song, C.; et al. Reversed-phase-reversed-phase liquidchromatography approach with high orthogonality for multidimen-sional separation of phosphopeptides. Anal. Chem. 2010, 82, 53−56.(23) Zhang, X.; et al. Multi-dimensional liquid chromatography inproteomics–a review. Anal. Chim. Acta 2010, 664, 101−113.(24) Zarei, M.; Sprenger, A.; Gretzmeier, C.; Dengjel, J.Combinatorial use of electrostatic repulsion-hydrophilic interactionchromatography (ERLIC) and strong cation exchange (SCX)chromatography for in-depth phosphoproteome analysis. J. ProteomeRes. 2012, 11, 4269−4276.(25) Villen, J.; et al. Large-scale phosphorylation analysis of mouseliver. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 1488−1493.(26) Alpert, A. J. Electrostatic repulsion hydrophilic interactionchromatography for isocratic separation of charged solutes andselective isolation of phosphopeptides. Anal. Chem. 2008, 80, 62−76.(27) Zarei, M.; Sprenger, A.; Metzger, F.; Gretzmeier, C.; Dengjel, J.Comparison of ERLIC-TiO2, HILIC-TiO2, and SCX-TiO2 for globalphosphoproteomics approaches. J. Proteome Res. 2011, 10, 3474−3483.(28) Dephoure, N.; Gygi, S. P. A solid phase extraction-basedplatform for rapid phosphoproteomic analysis. Methods 2011, 54,379−386.(29) Lin, Y.; et al. Sodium-deoxycholate-assisted tryptic digestion andidentification of proteolytically resistant proteins. Anal. Biochem. 2008,377, 259−266.(30) Leon, I. R.; et al. Quantitative assessment of in-solutiondigestion efficiency identifies optimal protocols for unbiased proteinanalysis. Mol. Cell. Proteomics 2013, 12, 2992−3005.



Dow

nloa

ded

by N

AT

L L

BR

Y O

F SE

RB

IA o

n Se

ptem

ber

11, 2

015

| http

://pu

bs.a

cs.o

rg

Pub

licat

ion

Dat

e (W

eb):

Nov

embe

r 4,

201

3 | d

oi: 1

0.10

21/p

r400

7969

http://pubs.acs.org

mailto:[email protected]



(31) Sugiyama, N.; et al. Phosphopeptide enrichment by aliphatichydroxy acid-modified metal oxide chromatography for nano-LC-MS/MS in proteomics applications. Mol. Cell Proteomics 2007, 6, 1103−1109.(32) Cox, J.; Mann, M. MaxQuant enables high peptide identificationrates, individualized p.p.b.-range mass accuracies and proteome-wideprotein quantification. Nat. Biotechnol. 2008, 26, 1367−1372.(33) Cox, J.; et al. Andromeda: a peptide search engine integratedinto the MaxQuant environment. J. Proteome Res. 2011, 10, 1794−1805.(34) Frese, C. K.; et al. Unambiguous phosphosite localization usingelectron-transfer/higher-energy collision dissociation (EThcD). J.Proteome Res. 2013, 12, 1520−1525.(35) Kim, M.-S.; Zhong, J.; Kandasamy, K.; Delanghe, B.; Pandey, A.Systematic evaluation of alternating CID and ETD fragmentation forphosphorylated peptides. Proteomics 2011, 11, 2568−2572.



Dow

nloa

ded

by N

AT

L L

BR

Y O

F SE

RB

IA o

n Se

ptem

ber

11, 2

015

| http

://pu

bs.a

cs.o

rg

Pub

licat

ion

Dat

e (W

eb):

Nov

embe

r 4,

201

3 | d

oi: 1

0.10

21/p

r400

7969

Rapid Combinatorial ERLIC−SCX Solid-Phase Extraction for In-Depth

Documents

Transcript of Rapid Combinatorial ERLIC−SCX Solid-Phase Extraction for In-Depth