Facilitating Phosphopeptide AnalysisUsing the Agilent HPLC Phosphochip

Abstract

The analysis of phosphopeptides can be particularly challenging since they are mostly

present at very low concentrations in a complex proteolytic sample. Selective enrich-

ment strategies have been introduced to capture phosphopeptides from the sample

prior to MS analysis, which can enhance detection capabilities. One of the more suc-

cessful phosphopeptide enrichment techniques is based on the selective capture of

phosphopeptides using titanium dioxide (TiO2) packing material. Although successful,

these enrichment strategies still face challenges with regards to automation, robust-

ness, reliability and reproducibility. The commercially available Agilent HPLC-Chip/MS

provides a platform that can integrate an enrichment column, analytical column, asso-

ciated connection capillaries, and a nanospray emitter directly on a single, small,

reusable microfluidic chip. It also provides an easy to use, robust and reliable

nanospray LC/MS platform when compared to conventional nanocolumn LC/MS

technology. Integration of enrichment columns with different stationary phase materi-

als on the chip can be used to selectively capture phosphopeptides during the sample

loading process. In this Application Note, we will describe the design and perfor-

mance of an HPLC-Chip with a hybrid TiO2 and C18 enrichment column for selective

phosphopeptide enrichment and analysis.

Authors

Reinout Raijmakers

Shabaz Mohammed

Albert J.R Heck

Utrecht University

Biomolecular Mass Spectrometry and

Proteomics Group

Bijvoet Center for Biomolecular

Research and Utrecht Institute for

Pharmaceutical Sciences

Padualaan 8,

3584 CH Utrecht

Netherlands

Application Note

Proteomics

2

Recent advances in HPLC-Chip/MStechnology have allowed the automa-tion of such a workflow, increasingease of use and confidence of analysis.The new microfluidic chip,Phosphochip, is a unique, re-usableHPLC-Chip specifically designed forphosphopeptide enrichment and analy-sis.6 The chip is a multilayer polyimidelaminate that contains an enrichmentsection with titanium dioxide (TiO2)beads flanked on both sides with C18reversed phase material. (Figure 1) Thethree sections in the sandwich enrich-

Introduction

Protein phosphorylation is one of themost important post-translational modi-fication (PTM) mechanisms for regulat-ing protein function in cells. Myriad bio-logical processes, including cell prolif-eration, migration, and apoptosisinvolve phosphorylation steps.1–3 Oneof the major efforts in proteomics isdevoted to the identification and under-standing of phosphoproteomes in cells.Nevertheless, comprehensive identifi-cation of sites of protein phosphoryla-tion remains a challenge, best left toexperienced proteomics experts. Multi-dimensional separation techniques cou-pling reversed phase (RP) with strongcation exchange (SCX) or a stronganion exchange (SAX) step prior to li-quid chromatography mass spectrome-try (LC/MS) analysis is necessary toachieve more sensitive results for com-plex mixtures. Furthermore, selectiveenrichment of phosphorylated proteinsand peptides is necessary to get betterphosphoproteome coverage. Commonlyused phosphopeptide enrichment tech-nologies are immobilized metal affinitychromatography (IMAC), metal oxidesof titanium, aluminum, or zirconium,and anti-phosphotyrosine antibodies.Although not completely overlapping,the titanium dioxide based approach isa particularly robust and automatablemethod with a high affinity for phos-phopeptides.4,5 However, multiple valveand micro-capillary connections createchallenges in reliability, robustness, andreproducibility. Automation of theenrichment step and analytical separa-tion is the best way to achieve run-to-run and lab-to-lab reproducibility.

ment column are separated from eachother by micro-fabricated frits. Thisenrichment section is connected to areversed phase separation column end-ing in an integrated electro-spray tip.Through the chip cube, the HPLC-Chip/MS interface, a micro valve isconnected directly with the chip sur-face. This creates a zero dead volumehigh pressure seal and the ultimateautomation for sample loading, desalt-ing, elution from trap column and gradient separation on the analyticalcolumn.

Packing material used:RP1: ZORBAX Extend 5 µm, 300ÅTiO2: 10 µm spheresRP2: ZORBAX Extend 5 µm, 300ÅAC : Reprosil-Pur C18-AQ 5 µm

Analytical Column

P1 P2

P3P6

Spray Tip

Enrichment Column

P4P5

Enrichment Column

A

B

Figure 1The unique design for Phosphochip. The Phosphochip (A) is a multi-layer microfluidic HPLC-Chip thatfeatures a sandwiched RP-TiO2-RP trapping column. (B) In the enrichment column, RP is illustrated inblack; TiO2 is illustrated in red; the figure on the left illustrates the loading position for peptide enrichmentand the figure on the right shows the separation of peptides through analytical column (please see nextpage on operation related to phosphopeptide enrichment and separation).

3

Experimental

Out-of-box Solution

The Phosphochip kit (G4240-62020) is acomplete solution for phosphopeptideanalysis on any Agilent HPLC-Chip/MSsetup. This could be a Chip/Ion Trap,Chip/QQQ or Chip/QTOF. Conditioningof the HPLC system is required in orderto saturate all potential phosphopeptidebinding sites in the system. A condi-tioning and standard mix is provided inthe reagent pack for this purpose. It isalso recommended to replace capillar-ies to and from the inline filter with polyaryletheretherketone (PEEK) tubingfrom the kit for operation at high pH.The unique sandwiched enrichment col-umn with RP1-TiO2-RP2 packings,allows the trapping and analysis of bothphosphorylated and non-phosphorylat-ed peptides in one experiment. Themicro valve connection on the chipallows the flow paths to be switched.The chip can be operated in two differ-ent modes for phosphopeptide enrich-ment. In the first mode, for phospho-peptide analysis only, most peptideswill be trapped on RP1. The elution ofpeptides with an organic plug pushingall through and onto the TiO2 phase,allows the enrichment of phosphopep-tides, with unbounded peptides going towaste. Phosphopeptides are then des-orbed with a special elution buffer andtrapped on RP2. Finally, the gradientelution of phosphopeptides occurs onthe analytical column. In dual modeanalysis, a gradient elution of non-phos-phorylated peptides is carried outinstead of an organic plug and flow isdirected through the analytical column.For more out-of-box operation of thePhosphochip, please refer to the User'sGuide (G4240-90005).

Materials and Methods

Reference protein mixtureThree proteins, bovine serum albuminand bovine alpha and beta casein, weredigested and combined to generate a

protein reference mixture. Each wasdissolved briefly in 1 M urea and 50 mM ammonium bicarbonate. Theproteins were reduced in 1 mM dithio-threitol (DTT) and alkylated in 2 mMiodoacetamide, followed by digestionwith trypsin overnight at a protein/pro-tease ratio of 50:1. The final mixturewas prepared for analysis by dilutingthe protein digests in 10% formic acidto a final concentration of 20 fmoL/µL.

Human cell lysateHuman U2OS osteosarcoma cells wereresuspended in lysis buffer with phos-phatase inhibitors (8 M urea, 50 mMNH4HCO3, 1 mM KF, 5 mM NaH2PO4, 1 mM Na2VO4), sonicated 2 times for 20 s each, then centrifuged at 13000 gand 4 ºC for 10 minutes to removeinsoluble material. Lysate proteinswere reduced with 1,4-dithiothreitol (10mM) and alkylated with iodoacetamidereagent (20 mM). The sample was thendiluted to 2 M of urea with lysis bufferlacking the urea, and 10 µg of trypsinwas added. The sample was incubatedfor 4 hours at 37 ºC. Next, the samplewas further diluted to 1 M of urea,another 10 ug of trypsin was added andthe sample was incubated overnight at 37 ºC. The resulting peptides were sep-arated by SCX chromatography and themain singly charged peptide fraction,known to also contain phosphopep-tides, was further analyzed by LC/MS/MS.

SCX fractionationSCX was performed using an Agilent1200 Series LC system with two C18Opti-Lynx (Optimized Technologies,Oregon City, OR) guard columns and apolysulfoethyl A SCX column (PolyLC,Columbia, MD; 200 mm × 2.1 mm innerdiameter, 5 µm, 200Å). The digestedcell lysate was dissolved in 0.05%formic acid and loaded onto the guardcolumn at 100 µL/min and subsequent-ly eluted onto the SCX column with

80% acetonitrile (ACN) and 0.05%formic acid. Peptides were fractionatedusing a multistage gradient of buffer A(5 mM KH2PO4, 30% ACN and 0.05% FA,pH 2.7) and buffer B (350 mM KCl, 5mM KH2PO4, 30% ACN and 0.05% FA,pH 2.7), and 1-minute fractions werecollected.

HPLC-Chip/MSAll LC/MS/MS experiments were per-formed using an Agilent 1200 SeriesHPLC-Chip/MS system connected to anAgilent 6520 Q-TOF mass spectrometeraccording to the Phosphochip User’sGuide. Samples were analyzed usingeither a "regular" C18 HPLC-Chip (15 cm, 75 µm Reprosil C18 analyticalcolumn) or a Phosphochip, operated indual mode, using 2-hour multi-step gra-dients of buffer A (0.6% HAc, 0.1% FA)and buffer B (0.6% HAc, 0.1% FA, 80%ACN). Per selected precursor (minimumintensity 2500), two MS2 spectra weregenerated with an acquisition time of1000 ms, with a 60 second dynamicexclusion.

Data analysisResulting MS2 spectra were searchedusing Spectrum Mill against the IPIhuman database (v3.37). MS2 spectrafrom the same precursor mass weremerged within a range of 1.4 m/z unitsand a retention time of 15 seconds.Variable modifications were the oxida-tion of methionine and the phosphoryla-tion of serine, threonine and tyrosine.Precursor and product mass tolerancewere both set at 50 ppm. Trypsin waschosen as the enzyme, with a maximumof 3 missed cleavages.

Results and discussion

Validation of phosphopeptideenrichmentAccording to the aforementioned proto-col, 50 fmol of a trypsin digested mix-ture of the bovine proteins serum albu-min (BSA), alpha-casein, and beta-

4

casein were applied, to validate thephosphopeptide enrichment of thePhosphochip. The majority of the pep-tides from the mixture were retained inthe flowthrough fraction of thePhosphochip, while only a small num-ber of peptides were detected in thePhosphochip elution. This is clear fromthe base peak chromatograms of thePhosphochip flowthrough and elutionfractions (Figure 2). When examiningspecific peptides, all the casein derivedphosphopeptides that were formedwere detected exclusively in the elutionfraction and not in the flowthrough(Figure 2). Comparison with a "regular"HPLC-Chip suggested near completebinding and recovery of the phospho-peptides with the TiO2 based chip.5 Themajority of the non-phosphorylated pep-tides were observed only in theflowthrough of the Phosphochip andnot in the elution (Figure 2).

Comparison between a regular HPLC-Chip and the PhosphochipSince biological samples are muchmore complex in composition and cancontain a large variety of different phos-phopeptides, prefractionation of thepeptides is required before the samplescan be efficiently analyzed byLC/MS/MS. One common approach forprefractionation is strong cationexchange (SCX), which can separatepeptides based on their net charge.5-8

Tryptic peptides most often have twobasic sites, the N-terminus of the pep-tide and the C-terminal lysine or argi-nine residue required for tryptic clea-vage, resulting in a 2+ net charge.Phosphopeptides, in contrast, willmainly have a single net charge as thephosphorylated residue, which is nega-tively charged compensates for onepositive charge.

Although this strategy enriches phos-phopeptides into a few SCX fractions,the same fractions can also containother peptides with a net single charge.

5 10 15 20 5

Time (min)

30 35 40 45

5 10 15 20 5

Time (min)

30 35 40 45

BPC

NAVPITPTLNR

YLYEIAR

FQpSEEQQQTEDELQDK

TVDEoxMpSTEVFTK

BPC

NAVPITPTLNR

YLYEIAR

FQpSEEQQQTEDELQDK

TVDEoxMpSTEVFTK

B

A Flowthrough

Elution

5

1

5

2

2

×105

×105

5

1

2

×105

×105

×105

×103

5

×103

×104

2

×104

Figure 2Enrichment of phosphopeptides from a standard protein mixture. (A) Base peak chromatograms (BPC)of the Phosphochip flowthrough and extracted ion chromatograms for two non-phosphorylated peptides(NAVPITPTLNR and YLYEIAR) and two phosphorylated peptides (FQpSEEQQQTEDELQDK andTVDEoxMpSTEVFTK) following analysis of a protein reference mixture. (B) identical chromatograms(note; equal y-axis) for the Phosphochip elution, revealing that the phosphopeptides are specificallyrecovered in the elution analysis.

5

This includes acetylated peptides andpeptides derived from protein C-termini,as well as the many other cellular com-pounds that can carry a single charge,all of which often are much more abun-dant than the phosphorylated peptidesof interest. Protein derived from ahuman osteosarcoma cell line (U2OS)was digested with trypsin to show theenrichment improvement that can beachieved on such samples using the

Figure 3Comparison of phosphopeptide analysis on a regular HPLC-Chip and a Phosphochip. (A) Total ion current chromatograms of the U2OS singly charged peptideSCX fraction analyzed on a regular HPLC-Chip (top, purple) and on a Phosphochip (flowthrough in red, elution in orange). (B) The number of unique phosphopep-tides identified with Spectrum Mill in the LC/MS/MS runs on the regular HPLC-Chip and in both the flowthrough and elution of the Phosphochip, as shown inpanel A. Indicated are the total number of MS2 spectra searched and the scoring and database used.

2

7

2

0

2

0

109

Unique PhosphopeptidesSpectrum Mill Score >9IPI human

Regular chip (TIC)11 phosphopeptides

Phosphochip Flowthrough (TIC)4 phosphopeptides

Phosphochip Elution (TIC)116 phosphopeptides

Regular chip PhosphochipFlowthrough

PhosphochipElution3969 MS2 spectra

4963 MS2 spectra

5309 MS2 spectra

5 10 15 20 5Time (min)

30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120

5 10 15 20 5Time (min)

30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120

5 10 15 20 5Time (min)

30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120

BA×107

1.2

×107

1.2

×107

1.2

Phosphochip. The resulting peptideswere separated by SCX. A single SCXfraction expected to contain the majori-ty of peptides with a net single chargewas analyzed both with a regular HPLCChip and with the Phosphochip (200 µg each).

As can be seen in Figure 3, the vastmajority of the sample was againretained in the flowthrough of thePhosphochip. It is noteworthy that a

large amount of the total sample (over90% of the total combined signal) isobserved in the flowthrough analysis onthe TiO2 chip. The subsequent analysisof the LC/MS/MS experiment usingthe Spectrum Mill search engine on thehuman IPI database allowed the identi-fication of 109 unique phosphopeptidesin the elution fraction of thePhosphochip with a Spectrum Millscore of at least 9. In contrast, only 11phosphopeptides were identified when

6

the same sample was analyzed with aregular C18 HPLC-Chip. As expected,the major non-phosphopeptide compo-nents hindered phosphopeptide identifi-cation for the sample without theenrichment step.

Large scale identification of phospho-peptides from human cellsA common use of TiO2 enrichment andSCX is global phosphoproteome charac-terizations. Often such experiments areperformed using milligrams of materialin order to gain access to phosphopep-tides present at low levels.7, 10, 11 Sinceit was clear that significantly morematerial could be analyzed by the TiO2

based HPLC-Chip, the sample amountwas increased by 5 times for easieridentification of phosphopeptides. Onemilligram of protein was digested andthe resulting peptides were separatedby SCX chromatography. The fullenriched singly charged fraction wasbrought on to a Phosphochip, allowing5 times more material to be analyzed.From the TIC chromatograms shown inFigure 4, it is clear that significantlymore material was observed in thePhosphochip elution. Equally apparentis the massive signal observed in theflowthrough analysis. Analysis of theLC/MS/MS results using Spectrum

Figure 4Large scale phosphopeptide enrichment from human cells. (A) Base peak chromatograms of the flowthrough (red) and elution (orange) from the large scaleanalysis of the U2OS singly charged peptide SCX fraction on the Phosphochip. (B) The number of unique phosphopeptides identified with Spectrum Mill fromboth the flowthrough and elution LC/MS/MS runs, shown in the panel A. Also indicated are the total number of MS2 spectra searched and the scoring anddatabase used.

Flowthrough (TIC)

Elution (TIC)

499

6

5

Flowthrough5194 MS2 spectra

Elution4803 MS2 spectra

A B

Unique PhosphopeptidesSpectrum Mill Score >9IPI human

5 10 15 20 5Time (min)

30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120

×107

2

5 10 15 20 5Time (min)

30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120

×107

2

Mill allowed the identification of over500 distinct phosphopeptides in theelution fraction using conservativeparameters, while only 12 phosphopep-tides were identified in the flowthroughanalysis.12 This scale of phosphopep-tide identifications is quite similar tomost recent global phosphoproteomeanalyses.7, 9

7

Phosphopeptide fragmentation andidentification from Phosphochip elu-tion fractionsThe combination of phosphopeptideenrichment with a Q-TOF mass spec-trometer has the advantage of highmass accuracy in fragmentation spec-tra, allowing the reliable identification

of phosphopeptides, particularly whenthe fragmentation of the peptides issuboptimal. Phosphopeptide fragmenta-tion spectra are generally difficult tointerpret as the loss of the phosphategroup is a common event during frag-mentation. This leads to low intensityof the product ions deriving backbone

Figure 5Fragmentation and identification of phosphopeptides from a human osteosarcoma cell line. Shown are the annotated fragmentation spectra of four phospho-peptides from the elution fraction of the Phosphochip analysis of net singly charged peptides from the human osteosarcoma cell line U2OS.

fragmentation, which are required foridentification. Figure 5 shows variousphosphopeptide fragmentation spectrawith their unambiguous assignment tophosphorylated human peptides, asobtained for peptides originating fromthe U2OS osteosarcoma cell line.

A

B

C

www.agilent.com/chem/hplc-chip

© Agilent Technologies, Inc., 2009June 10, 2009Publication Number 5990-4098EN

Conclusion

These experiments demonstrated theefficiency of phosphopeptide enrich-ment and analysis on a microfluidicHPLC-Chip, the Phosphochip. ThePhosphochip workflow combinesAgilent 1200 Nano and Capillary LC,HPLC-Chip/MS interface, and anyAgilent 6000 Series mass spectrometer.This ultimate ease-of-use tool will helpresearchers studying phosphorylationachieve an increase in unique phospho-peptide identification from complexmixtures.

References

1.O. N. Jensen, “Interpreting the ProteinLanguage Using Proteomics,” Nat. Rev.Mol. Cell Biol. 2006, 7 (6), 391-403.

2.G. Manning, et al. “The Protein KinaseComplement of the Human Genome,”Science 2002, 298 (5600), 1912.

3.M. Mann, O. N. Jensen, “ProteomicAnalysis of Post-TranslationalModifications,” Nat. Biotechnol. 2003,21 (3), 255-261.

4.M. W. Pinkse, et al., “Selective Isolationat the Femtomole Level ofPhosphopeptides from ProteolyticDigests Using 2D-NanoLC-ESI-MS/MSand Titanium Oxide Precolumns,” AnalChem, 2004. 76(14): 3935-43.

5.M. W. Pinkse, et al., “Highly Robust,Automated, and Sensitive Online TiO2-Based Phosphoproteomics Applied toStudy Endogenous Phosphorylation inDrosophila Melanogaster,” J ProteomeRes, 2008. 7(2): 687-97.

6.S. Mohammed, et al., “Chip-BasedEnrichment and NanoLC/MS/MSAnalysis of Phosphopeptides fromWhole Lysates,” J Proteome Res, 2008.7(4): 1565-71.

7.S. Lemeer, et al., “ComparativePhosphoproteomics of ZebrafishFyn/Yes Morpholino KnockdownEmbryos,” Mol Cell Proteomics, 2008.7(11): 2176-87.

8.J. Peng, J. E. Elias, C. C. Thoreen, L. J.Licklider, and S. P. Gygi, “Evaluation ofmultidimensional chromatography cou-pled with tandem mass spectrometry(LC/LC-MS/MS) for large-scale proteinanalysis: the yeast proteome” JProteome Res, 2003. 2(1): 43–50.

9.J. Villen, and S. P. Gygi, “TheSCX/IMAC Enrichment Approach forGlobal Phosphorylation Analysis byMass Spectrometry.” Nat Protoc, 2008.3(10): 1630-8.

Ordering Information

Product number

Phosphochip kit G4240-62020Includes Phosphochip, Reagent Pack (400-600, with ready-to-use Elution Buffer and Regeneration Mix and lyophilized Conditioning/Standard Mix)

Phosphochip only G4240-62021(Orderable beginning of 2010)

Reagent Pack only 400-600

PEEK sample transfer capillary (25 µm x 100 cm) G4240-87309

PEEK capillary for use with micro inline filter (25 µm x 10 cm) G4240-87310

Chip care and use sheet G4240-90001

Phosphochip User's Guide G4240-90005

10.J. V. Olsen, et al., “Global, In Vivo, andSite-Specific Phosphorylation Dynamicsin Signaling Networks,” Cell, 2006.127(3): 635-48.

11.J. Villen, et al., “Large-ScalePhosphorylation Analysis of MouseLiver,” Proc Natl Acad Sci U S A, 2007.104(5): 1488-93.

12.Raijmakers et al., “Exploring the humanleukocyte phosphoproteome using amicrofluidic RP-TiO2-RP HPLCPhosphochip coupled to a Q-ToF massspectrometer” submitted 2009.