Techniques for Phosphopeptide Enrichment Prior to Mass Spec Anal

TECHNIQUES FOR PHOSPHOPEPTIDE ENRICHMENT PRIOR TOANALYSIS BY MASS SPECTROMETRY

Jamie D. Dunn,1 Gavin E. Reid,1,2 and Merlin L. Bruening1*1Department of Chemistry, Michigan State University,East Lansing, Michigan 488242Department of Biochemistry and Molecular Biology,Michigan State University, East Lansing, Michigan 48824

Received 2 July 2008; received (revised) 17 November 2008; accepted 17 November 2008

Published online 4 March 2009 in Wiley InterScience (www.interscience.wiley.com) DOI 10.1002/mas.20219

Mass spectrometry is the tool of choice to investigate proteinphosphorylation, which plays a vital role in cell regulationand diseases such as cancer. However, low abundances ofphosphopeptides and low degrees of phosphorylation typicallynecessitate isolation and concentration of phosphopeptides priorto MS analysis. This review discusses the enrichment ofphosphopeptides with immobilized metal affinity chromato-graphy, reversible covalent binding, and metal oxide affinitychromatography. Capture of phosphopeptides on TiO2 seemsespecially promising in terms of selectivity and recovery, but thesuccess of all methods depends on careful selection of binding,washing, and elution solutions. Enrichment techniques arecomplementary, such that a combination of methods greatlyenhances the number of phosphopeptides isolated from complexsamples. Development of a standard series of phosphopeptides ina highly complex mixture of digested proteins would greatly aidthe comparison of different enrichment methods. Phosphopeptidebinding to magnetic beads and on-plate isolation prior toMALDI-MS are emerging as convenient methods for purificationof small (mL) samples. On-plate enrichment can yield >70%recoveries of phosphopeptides in mixtures of a few digestedproteins and can avoid sample-handling steps, but this techniqueis likely limited to relatively simple samples such as immuno-precipitates. With recent advances in enrichment techniques inhand, MS analysis should provide important insights intophosphorylation pathways. # 2009 Wiley Periodicals, Inc.,Mass Spec Rev 29:29–54, 2010Keywords: phosphorylation; enrichment; IMAC; titaniumdioxide; magnetic beads; metal oxides

I. ABBREVIATIONS

ACTH adrenocorticotropic hormoneBSA bovine serum albuminCHAPS 3-[(3-cholamidopropyl)dimethylammonio]-1-

propanesulfonatea-CHCA a-cyano-4-hydroxycinnamic acid matrix2,5-DHB 2,5-dihydroxybenzoic acid

DIPEA N,N0-diisopropylethylamineDTT dithiothreitolEDC N,N0-dimethylaminopropyl ethyl carbodiimideEDTA ethylenediaminetetraacetic acidESI electrospray ionizationFT-ICR Fourier transform ion cyclotron resonanceGLYMO 3-glycidoxypropyltrimethoxysilaneGMA glycidyl methacrylatecGMP cyclic guanosine monophosphateHEMA 2-hydroxyethyl methacrylateb-HPA b-hydroxypropanoic acidIDA iminodiacetateIMAC immobilized metal affinity chromatographyiTRAQ isobaric tag for relative and absolute protein

quantitationLC liquid chromatographyMALDI matrix-assisted laser desorption/ionizationMOAC metal oxide affinity chromatographyMS mass spectrometryNP nanoparticleNTA nitrilotriacetatePAA poly(acrylic acid)PBS phosphate-buffered salinePEI polyethyleneiminePHEMA poly(2-hydroxyethyl methacrylate)PKG cGMP-dependent kinasePySSPy 2,20-dithiopyridineSAMs self-assembled monolayersSCX strong cation exchangeSALDI surface-assisted laser desorption/ionizationSELDI surface-enhanced laser desorption/ionizationSILAC stable isotope-labeling of amino acids in cell cultureTEOS tetraethyl orthosilicateTFA trifluoroacetic acidTHAP 20,40,60-trihydroxyacetophenoneTMSPED N-[3-(trimethoxysilyl)propyl]ethylenediamineTOF time-of-flight

II. INTRODUCTION

Protein phosphorylation is one of the most important mecha-nisms to regulate cellular processes (Graves & Krebs, 1999) suchas gene expression and membrane transport (Krebs & Beavo,1979; Krebs, 1983; Hunter, 2000; Pawson & Nash, 2000;

Mass Spectrometry Reviews, 2010, 29, 29– 54# 2009 by Wiley Periodicals, Inc.

————*Correspondence to: Merlin L. Bruening, Department of Chemistry,

Michigan State University, East Lansing, MI 48824.

E-mail: [email protected]

Whitmarsh & Davis, 2000; Adams, 2001; Johnson & Lewis,2001; Simpson, 2003). In a number of regulatory pathways,protein kinases and phosphatases regulate the post-translationalphosphorylation status of serine, threonine, and tyrosine residues(Adams, 2001; Johnson & Hunter, 2005), and disruption of theseregulatory pathways can sometimes contribute to diseasessuch as cancer (Cohen, 2001; Lim, 2005). Thus, identificationof phosphorylation sites might be vital to develop newpharmaceutical targets and to understand disease states (Fischer& Krebs, 1955; Cohen, 2001; Lim, 2005; Yang et al., 2006; Yu,Issaq, & Veenstra, 2007). Mass spectrometry is currently themethod of choice to detect changes in protein phosphorylationand to identify the position of specific phosphorylation events(Mann et al., 2002).

However, even with recent advances in mass spectrometryinstrumentation, the detection and identification of phosphor-ylation sites is challenging. Frequently, because both the amountof phosphorylated protein in eukaryotic cells and the degree ofphosphorylation are relatively low (Aebersold & Goodlett, 2001;Mann et al., 2002; Simpson, 2003) highly sensitive methods areneeded. Additionally, two studies suggested that ionizationefficiencies and MS signals of phosphorylated peptides arelower than those of their nonphosphorylated analogues, and thisfactor could make it difficult to detect phosphorylated species inthe presence of an abundance of nonphosphorylated peptides(Craig et al., 1994; Liao et al., 1994). For matrix-assisted laserdesorption/ionization mass spectrometry (MALDI-MS), desorp-tion/ionization efficiencies for phosphopeptides were reportedto be an order of magnitude lower than those of theirnonphosphorylated counterparts, and ionization of phosphory-lated peptides can become more difficult as the number ofphosphorylation sites increases. More recently, the validity ofthese results was tested with high-performance LC-MS with avariety of synthetic phosphopeptides and their nonphosphory-lated counterparts spiked in tryptic protein digests (Steen et al.,2006). This study concluded that the difficulty to detectphosphopeptides with LC-MS stems from the low abundanceof the species, not low ionization efficiencies. However, thesynthetic peptides examined contained many basic amino acidresidues such as arginine (Steen et al., 2006), which generallyallow for better MS detection (Krause, Wenschuh, & Jungblut,1999; Clipston, Jai-nhuknan, & Cassady, 2003). Additionally, thepeptides were separated with LC prior to analysis by MS, sofurther studies with mixtures of peptides might be needed. Inany case, the detection and identification of phosphorylatedspecies is challenging because of low abundances, so enrichmenttechniques are typically employed prior to analysis.

This article reviews recently developed methods for theselective capture and elution of phosphopeptides prior toanalysis, along with adaptations of these methods for bothenrichment using magnetic beads and on-plate purification priorto MALDI-MS. Although there are numerous reports on theapplication of mass spectrometry for identifying phosphorylationsites, the focus here is on enrichment techniques, not applica-tions. The three major methods to capture phosphopeptidesinclude immobilized metal affinity chromatography (IMAC),reversible covalent binding, and metal oxide affinity chromatog-raphy (MOAC), and these methods are discussed in Section II.Avidin affinity chromatography has also been described for

phosphopeptide capture (Adamczyk, Gebler, & Wu, 2001; Oda,Nagasu, & Chait, 2001; Goshe et al., 2002), but this techniquerequires a number of steps and relatively large sample quantitiesand is not described here. Strong cation exchange (SCX)chromatography is another recent alternative employed forphosphopeptide enrichment, (Ballif et al., 2004; Beausoleil et al.,2004; Lim & Kassel, 2006; Olsen et al., 2006; Trinidad et al.,2006; Wu et al., 2007) but this technique has also been reviewedrecently (Gafken & Lampe, 2006; Yu, Issaq, & Veenstra, 2007).SCX is frequently employed in combination with IMAC, MOAC,and covalent techniques.

Finally, immunoprecipitation is capable of enrichingtyrosine-phosphorylated proteins and peptides and greatly aidsin the identification of sites of tyrosine phosphorylation (Zhang &Neubert, 2006; Schmidt, Schweikart, & Andersson, 2007). Inhyperphosphorylated Jurkat cells, Rush and co-workers usedphosphotyrosine-specific antibodies to enrich tryptic tyrosinephosphopeptides and identify 194 phosphotyrosine sites (Rushet al., 2005). Some studies employed antibodies to specific motifsin phosphothreonine and phosphoserine peptides (Grønborget al., 2002; Matsuoka et al., 2007), but antibody-based isolationof phosphoserine and phosphothreonine peptides is not yet ageneral technique (Kristjansdottir et al., 2008; Sopko &Andrews, 2008) so this method is not discussed here.

Because recent reviews focused on IMAC and covalentcapture (Reinders & Sickmann, 2005; Morandell et al., 2006;Schmelzle & White, 2006; D’Ambrosio et al., 2007; Yu, Issaq, &Veenstra, 2007), we will present these methods relatively brieflyand will describe advances in MOAC in more detail. Section IIIof this review describes the adaptation of IMAC and MOAC forselective capture on magnetic beads and isolation of phospho-peptides directly on MALDI plates. Although these platformsmay not be well-suited to identify all of the phosphopeptides in acomplex sample such as a cell extract, they might prove usefulto rapidly analyze small volumes of dilute samples. Finally,Section IV presents a brief perspective on phosphopeptideenrichment.

III. PHOSPHOPEPTIDE CAPTURE AND ELUTION

A. Immobilized Metal Affinity Chromatography

Immobilized metal affinity chromatography (IMAC) relies onthe interactions between metal-ligand complexes and specificfunctional groups such as phosphates. The enrichment procedureincludes selective binding to the resin, a rinse to removeimpurities, and elution to yield a concentrated solution of thespecies of interest, as shown in Figure 1. Although IMAC is lessspecific than other affinity methods, this lower specificity makesit applicable to a wide range of separations, including enrichmentof phosphorylated species. IMAC was introduced in 1975 (Porathet al., 1975), and has historically been the method mostcommonly used to isolate phosphopeptides prior to analysis byMS (Neville et al., 1997; Li & Dass, 1999).

Iminodiacetate (IDA) and nitrilotriacetate (NTA), shown inFigure 2, are the prototypical metal-binding ligands employed inIMAC stationary phases. NTA was introduced to IMAC in 1987(Hochuli, Dobeli, & Schacher, 1987), and the Fe(III)–NTA

& DUNN, REID, AND BRUENING

30 Mass Spectrometry Reviews DOI 10.1002/mas

complex is probably the complex most frequently employed toenrich phosphopeptides, although Ga(III), Zr(IV), and Al(III)complexes have also been investigated for their affinity towardsphosphorylated species (Posewitz & Tempst, 1999; Nuhse et al.,2003). Alternative metal-ligand complexes of Zr(IV)-phospho-nate immobilized on various stationary phases or supports havebeen recently employed for phosphopeptide enrichment byseveral groups (Zhou et al., 2006; Dong et al., 2007; Feng et al.,2007; Yu, Issaq, & Veenstra, 2007; Wei et al., 2008).

There are a number of challenges associated with usingIMAC. First, because the metal ions are not covalently bound tothe substrate, there is a possibility to leach these ions from thecolumn during the enrichment steps, which might lead to loss ofphosphopeptides or to a contamination of peptides with metalions. However, thorough washing of the column before use and ajudicious selection of binding ligands can minimize theseproblems. For example, IDA is a tridentate ligand, whereasNTA is a tetradentate ligand; hence, the latter binds more stronglyto the metal ion to better prevent leaching (Hochuli, Dobeli, &Schacher, 1987; Holmes & Schiller, 1997).

A second challenge is nonspecific binding of peptides thatcontain the acidic residues glutamic and aspartic acid. A fewreports suggested that increasing the ionic strength of the loadingor a rinsing solution can help minimize electrostatic interactionsbetween acidic residues and metal-ion complexes (Andersson &Porath, 1986; Holmes & Schiller, 1997; Kagedal, 1998; Ficarroet al., 2002). However, another study did not see a significantreduction in nonspecific adsorption by incorporating salts into theenrichment protocol (Ndassa et al., 2006). To better overcomethis unwanted binding, the carboxyl groups of amino acidresidues can be converted to methyl esters (Ficarro et al., 2002).Peptides are typically esterified by reaction with acetyl chlorideFIGURE 2. Structures of (a) IDA and (b) NTA.

FIGURE 1. Schematic diagram of phosphopeptide isolation by IMAC or MOAC. The column packing is

different for the two techniques. [Color figure can be viewed in the online issue, which is available at

www.interscience.wiley.com.]

TECHNIQUES FOR PHOSPHOPEPTIDE ENRICHMENT &

Mass Spectrometry Reviews DOI 10.1002/mas 31

in an excess of methanol (methanolic HCl). More recently,thionyl chloride in methanol was used for formation of the methylesters of phosphopeptides (Moser & White, 2006; Ndassa et al.,2006). The authors reported that this reagent showed a higherefficiency for the conversion of carboxylic acid groups to methylesters than did methanolic HCl. Unfortunately, incompleteesterification of phosphopeptides will increase chemical noise,and sample losses during lyophilization of the esterified peptidesmay be around 20% (Stewart, Thomson, & Figeys, 2001).Bodenmiller and co-workers examined the effect of methylesterification on the identification of phosphopeptides that wereisolated from a cytosolic fraction of Drosophila melanogasterKc167 cells (Bodenmiller et al., 2007a). Methyl esterificationprior to IMAC allowed identification of 199 peptides, whereasIMAC prior to methyl esterification permitted identification of193 peptides. Interestingly, the overlap of the peptides identifiedby the two protocols was only approximately 30%.

Nonspecific binding to Ga(III)-IMAC resins can also bereduced by using the endoproteinase glu-C rather than trypsin forprotein digestion (Seeley, Riggs, & Regnier, 2005). Glu-Ccleaves the protein at the C-terminus of glutamic and asparticacid residues so that only one acidic residue will be present in thepeptide chains. Hence, glu-C digestion should alleviate essen-tially all nonspecific binding due to interactions between IMACresins and multiple acidic residues.

Another challenge that has been reported in the use of IMACis that the technique is more specific for multiply phosphorylatedpeptides than monophosphorylated peptides because multiplyphosphorylated peptides bind more strongly to the IMAC resin(Ficarro et al., 2002; Nuhse et al., 2003; Nousiainen et al., 2006;Jensen & Larsen, 2007). Thingholm and co-workers utilized thestrong binding of multiply phosphorylated peptides to IMACresins to increase the number of phosphorylated peptidesdetected from lysates of human mesenchymal stem cells(Thingholm et al., 2008). They first eluted monophosphopeptidesfrom the IMAC column using 1% TFA and purified both thecolumn flow-through and the eluent a second time using TiO2.Elution using NH4OH then yielded a fraction rich in multiplyphosphorylated peptides. Overall, this combination of enrich-ment techniques facilitated identification of 492 phosphorylatedpeptides, including 186 multiply phosphorylated peptides,whereas the use of just TiO2 enrichment revealed only286 phosphopeptides, 54 of which were multiply phosphory-lated. Thus, the combination of IMAC with different eluentsis particularly useful for detecting multiply phosphorylatedspecies.

The specificity of IMAC for binding multiply phosphory-lated peptides depends on variables such as the affinity ligand, thebinding capacity of the support material, and the binding, rinsing,and elution conditions. Recently, the use of a high-capacityIMAC material in conjunction with an optimized binding andrinsing buffer (33:33:33 acetonitrile/methanol/aqueous 0.1%acetic acid) led to enhanced phosphopeptide recovery anduniform LC-MS detection of multiply and singly phosphorylatedpeptides (Ndassa et al., 2006). Peptides that contained acidresidues were converted to methyl esters in this study. If theIMAC column does not have a large enough capacity for all of thephosphorylated peptides in the mixture, then the multiplyphosphorylated peptides will bind preferentially over the mono-

phosphorylated species. Reducing the concentration of aceticacid from 1% to 0.1% yielded an increase in the numberof phosphorylated peptides detected, and the recovery ofphosphorylated peptides from a cell lysate doubled with33:33:33 acetonitrile/methanol/aqueous 0.1% acetic acid as theloading and washing solutions instead of 25:75 acetonitrile/aqueous 1% acetic acid that contained 100 mM NaCl. In otherexperiments, the use of 100 mM NaCl had the greatest effect onmonophosphorylated peptide recoveries, which decreased byhalf when NaCl was present. Another group implementedtrifluoroacetic acid (TFA) rather than acetic acid, and a highcontent of acetonitrile in the loading and rinsing solutions toreduce nonspecific adsorption from acidic and hydrophobicpeptides (Kokubu et al., 2005). Effective loading and washingsolutions for a pipette tip loaded with IMAC resin (Phos-Selectgel from Sigma, St. Louis, MO) contained 40–60% acetonitrileand 0.1–1% TFA.

In some cases, IMAC columns were applied to relativelycomplex samples. Giorgianni and co-workers used IMAC alongwith both a C18 column and a POROS Oligo R3 column to collectphosphopeptide fractions from prostate cancer cells (Giorgianniet al., 2007). This study identified 137 phosphorylation sites in 81phosphoproteins. Related work used a combination of isoelectricfocusing and IMAC to identify 50 phosphorylation sites in 26proteins obtained from primary pituitary tissue (Beranova-Giorgianni et al., 2006). Thus, IMAC can be applied to relativelycomplex samples, but it is typically used in combinationwith other separation techniques. As mentioned above, theuse of multiple fractions from IMAC allowed identification of492 phosphopeptides in human stem cells (Thingholm et al.,2008).

B. Reversible Covalent Binding

In principle, covalent binding might offer higher enrichmentselectivities than IMAC, but typical covalent proceduresfrequently employ a number of steps. In 2001, a multistep,solid-phase enrichment technique that can be applied tophosphotyrosine peptides in addition to phosphoserine- andphosphothreonine-containing peptides was developed (Zhou,Watts, & Aebersold, 2001). The method is portrayed in Figure 3.Prior to tryptic digestion, any cysteine residues of the proteinmixture are reduced and alkylated. After digestion, the aminogroups of the peptides are first protected with t-butyl-dicarbonate(tBoc), and the carboxylate and phosphate groups areconverted to amides and phosphoramides through a carbodiimide(N,N0-dimethylaminopropyl ethyl carbodiimide, EDC)-medi-ated reaction with ethanolamine. The phosphate group isregenerated by acid hydrolysis (10% TFA), and cystamine isattached to the phosphate group via a carbodiimide-catalyzedcondensation. Finally, the attached cystamine is reduced withdithiothreitol (DTT), and the free sulfhydryl group is reacted withiodoacetyl groups immobilized on glass beads. After rinsing thebeads with 2 M NaCl, methanol, and water to removenonphosphorylated peptides, the phosphopeptides are cleavedfrom the surface and the tBoc protecting groups are removed with100% TFA. Using this elaborate enrichment strategy for theanalysis of the phosphotyrosine peptide TTHpYGSLPQK in a



digest of phosphorylated myelin basic protein, recovery was only20% (Zhou, Watts, & Aebersold, 2001).

b-Elimination chemistry provides a considerably simplermethod to enrich phosphopeptides. In this chemistry, strongbases such as NaOH or Ba(OH)2 are used to cleave thephosphoester bonds of phosphoserine and phosphothreonine toform the respective dehydroalanine or dehydroaminobutyric acidanalogs, which can react with different nucleophiles, such asthiol, amine, or alcohol groups (Fig. 4). These reactions were

used to cleave the phosphate ester of a-casein phosphopeptidesfrom a tryptic digest (cysteines were oxidized to cysteic acid priorto digestion), and these peptides were subsequently reactedwith propanedithiol and covalently bound to a solid supportderivatized with reactive dithiopyridine groups (Thaler et al.,2003). After the resin was rinsed thoroughly, the bound peptideswere simply cleaved using DTT, and the free thiol groups werealkylated. The sample was desalted and analyzed using MALDI-TOF-MS. However, when combined with MALDI-MS, this

FIGURE 4. Reversible solid-phase enrichment of phosphoserine-containing peptides using

b-elimination and Michael addition followed by enrichment on a dithiopyridino-modified resin

(pro edure from Thaler et al., 2003). Enrichment can also be applied to phosphothreonine-containing

peptides.

FIGURE 3. Solid-phase enrichment using carbodiimide condensation and glass beads derivatized with

iodoacetyl groups. In the final step, the phosphopeptide is cleaved from the bead and the tBoc protecting

group is removed using 100% TFA, which is not shown in this figure (procedure from Zhou, Watts, &

Aebersold, 2001).



method revealed only 2 tryptic phosphopeptides from a 20-mg(�820 pmol) a-casein digest. a-Casein contains a-S1 and a-S2protein forms, and complete tryptic digestion of these proteinsshould give a total of 9 phosphopeptides.

Another covalent enrichment technique employed an a-diazo functionalized resin to reversibly and covalently bind thephosphate group of phosphopeptides (Lansdell & Tepe, 2004).Because b-elimination or another technique is not required tochemically derivatize the phosphopeptide prior to solid-phaseenrichment, this technique can be applied to phosphorylatedserine, threonine, and tyrosine peptides. Figure 5 shows theimmobilization procedure. To prevent carboxylate groups ofthe peptides from covalently binding to the resin, these groupswere first converted to methyl esters. The authors used thea-diazo functionalized resin to enrich 500 fmol of phosphory-lated angiotensin II (DRVpYIHPF) from a mixture of threenonphosphorylated peptides. After the peptide mixture wasincubated with the a-diazo resin, the nonphosphopeptideswere rinsed away, and the immobilized phosphopeptide wascleaved with either TFA or NH4OH. The resulting peptides wereanalyzed with MALDI-TOF-MS, and the mass spectrum showeda single peak, which was due to phosphorylated angiotensin (m/z1127).

Figure 6 shows the immobilization of methyl-esterifiedphosphopeptides onto an amino-terminated dendrimer with EDCcoupling (Tao et al., 2005). This one-pot chemistry avoids theneed to protect the amino groups on the N-terminus and lysine orarginine residues, presumably because of the large excess ofamine groups on the dendrimers. Once the phosphoprotein digesthad incubated with the dendrimer for 10 hr, the dendrimer wasrinsed with 3 M NaCl, 30% methanol, and water, and thephosphopeptides were liberated (10% TFA, 30 min) to cleave thephosphoramidate bonds. This technique was employed toanalyze digests of b-casein, and the phosphopeptide recoverywas greater than 35%. The same group also showed thatphosphopeptides activated with EDC also react with excesscystamine to form phosphoramidate bonds. The modifiedphosphopeptides can then be immobilized via reaction ofreduced cystamine thiol groups with maleimide functionalitiesin pore glass. Finally, cleavage of the phosphoramidate bondunder acidic conditions liberates the peptide. Using this method,the recovery of 100 fmol of phosphorylated angiotensin from a500 pmol BSA digest was approximately 40% (Bodenmilleret al., 2007b).

Warthaka and co-workers developed a reversible solid-phase enrichment technique based on oxidation-reduction

FIGURE 5. Reversible solid-phase enrichment using ana-diazo resin (Lansdell & Tepe, 2004). Cleavage of

the phosphopeptide can be accomplished by using either trifluoroacetic acid (TFA) or NH4OH. Note that

when NH4OH is used, the methyl ester is hydrolyzed back to the original phosphopeptide.



condensation of phosphopeptides with glycine-derivatizedWang resin as shown in Figure 7 (Warthaka, Karwowska-Desaulniers, & Pflum, 2006). In the first step, carboxylic acid-containing residues were protected via methyl esterification.Next, the methylated phosphopeptides were covalentlycoupled to the glycine Wang resin in the presence oftriphenylphosphine (PPh3), 2,20-dithiopyridine (PySSPy), andN,N0-diisopropylethylamine (DIPEA). The phosphopeptide wasbound to the glycine-derivatized resin via a phosphoramidatebond, which is cleavable with TFA. Subsequently, the resinwas washed and the phosphopeptides were eluted with 95%TFA and analyzed with MALDI-TOF-MS. By using this

enrichment strategy, the recovery of a monophosphopeptidefrom a b-casein digest was �37% (Warthaka, Karwowska-Desaulniers, & Pflum, 2006), which was comparable to that of thecarbodiimide solid-phase enrichment technique mentionedabove (Tao et al., 2005).

C. Metal Oxide Affinity Chromatography

The use of MOAC for phosphopeptide enrichment has grownrapidly in the past 4 years because of high recoveries andselectivities, even without methyl esterification of peptides.

FIGURE 6. Three-step solid phase enrichment using carbodiimide condensation (procedure from Tao

et al., 2005).

FIGURE 7. Reversible solid-phase enrichment of protected phosphopeptides using an oxidation-reduction

condensation reaction (procedure from Warthaka, Karwowska-Desaulniers, & Pflum, 2006). The amino

terminus was acetylated using acetic anhydride.



Additionally, the metal oxides are often more stable thantraditional silica-based stationary phases. ZrO2, for example, isstable at temperatures up to 2008C and over a pH range from 1 to14 (Nawrocki et al., 2004). Metal oxides that have been appliedtowards selective phosphopeptide enrichment include titaniumdioxide (TiO2), zirconium dioxide (ZrO2), aluminum oxide(Al2O3), and niobium oxide (Nb2O5). Aluminum hydroxide hasalso been used. Below we discuss phosphopeptide enrichmentstudies with each of these materials.

1. Enrichment with Titanium Dioxide

Currently, titanium dioxide is the most popular metal oxide resinused as a selective affinity support to capture phosphorylatedcompounds, including peptides (Ikeguchi & Nakamura, 1997,2000; Pinkse et al., 2004; Sano & Nakamura, 2004a,b; Larsenet al., 2005). At acidic pH, TiO2 has a positively charged surface(Kosmulski, 2002) that selectively adsorbs phosphorylatedspecies and exhibits outstanding enrichment behavior forphosphopeptides (Fig. 1). Notably, when coupled with appro-priate solutions for sample loading, column rinsing, and peptideelution, TiO2 is highly selective to preferentially bind phospho-peptides over acidic peptides (Larsen et al., 2005; Thingholmet al., 2006; Jensen & Larsen, 2007; Sugiyama et al., 2007).

TiO2 precolumns and 2D-nano-LC-ESI-MS/MS wereutilized to analyze a synthetic phosphopeptide as well asproteolytic digests of cGMP-dependent protein kinase (Pinkseet al., 2004). The protocol called for a phosphopeptide loadingsolution in 0.1 M acetic acid, a rinsing solution of 0.1 M aceticacid in 80% acetonitrile, and an ammonium bicarbonateeluent (pH 9.0). When an equimolar mixture of 125 fmol ofRKIpSASEF, a synthetic phosphorylated peptide derived fromcGMP-dependent protein kinase (PKG), and 125 fmol of itsnonphosphorylated counterpart were enriched with the TiO2

precolumn, the percent recovery of the RKIpSASEF was above90%. Additionally, 11 phosphorylated peptides were recoveredfrom the analysis of a PKG tryptic digest. However, non-phosphorylated peptides were also retained on the titaniumdioxide precolumn. The authors thought that the acidic nature ofthese nonphosphorylated peptides caused them to have an affinityfor TiO2, so they compared the affinity of the methylated andnonmethylated forms of [Glu1]-fibrinopeptide B (EGVND-NEEGFFSAR) for titania. Their studies showed that 98% ofthe nonmethylated peptide was retained on the TiO2 precolumnunder the phosphopeptide loading and washing procedure, andeluted under basic conditions. In contrast, the methylated formappeared to have very little affinity for TiO2.

In other developments with TiO2, the modification of theloading buffer and rinsing solution with 0.1% TFA and 2,5-dihydroxybenzoic acid (2,5-DHB) rather than 0.1 M acetic acidimproved the detection of phosphorylated peptides (Larsen et al.,2005). Hart and co-workers first introduced 2,5-DHB as an eluentfor IMAC (Hart et al., 2002). The pH values of 0.1% TFA and0.1 M acetic acid solutions are 1.9 and 2.7, respectively; hence,the TFA solution protonates acidic residues and preventsadsorption of nonphosphorylated peptides to TiO2. Moreover,an NH4OH eluent at pH 10.5 afforded a high recovery ofphosphopeptides. The pKa and pKb values for titania are 4.4 and

7.7, respectively (Koizumi & Taya, 2002), so at high pH, thetitania should become negatively charged to allow phosphopep-tide elution. (The first and second pKa values of phosphoserineand phosphothreonine are <1.7 and 6, respectively; Vogel, 1989;Xie, Jiang, & Ben-Amotz, 2005.) Four additional phosphopep-tides were observed from an a-casein digest with NH4OH ratherthan pH 9.0 ammonium bicarbonate as the eluent. Remarkably,when 2,5-DHB was used in the binding and rinsing solutionand NH4OH, pH 10.5 was the eluent, 20 phosphorylatedpeptides were detected in the MALDI mass spectrum of a500 fmol a-casein digest, whereas virtually no signals due tononphosphorylated peptides were observed (Larsen et al., 2005).Larsen and co-workers also optimized conditions when theyanalyzed a more complex digest mixture that containedequimolar amounts (500 fmol) of three nonphosphorylatedproteins (bovine serum albumin (BSA), b-lactoglobulin, andcarbonic anhydrase) and three phosphoproteins (b-casein,a-casein, and ovalbumin). Using enrichment with TiO2 alongwith the loading solutions and eluents mentioned above, theywere able to recover 18 phosphorylated peptides, whereas themajority of peaks due to nonphosphorylated peptides werevirtually eliminated. It should be noted that after the TiO2

column, these studies typically used a Poros Oligo R3 micro-column for sample desalting and concentration. Occasionally,phosphopeptides were not retained on the Oligo material andrequired further purification on a graphite microcolumn. Anenrichment comparison was made between this titania materialand Fe(III)–NTA–IMAC beads (PHOS-select, Sigma) with amixture of the digested proteins described above. At 10:1 and50:1 molar ratios of nonphosphoproteins to phosphoproteins intryptic digests, the use of the commercial PHOS-select IMACbeads (Sigma) yielded fewer phosphorylated peptide signals andmore peaks due to nonphosphopeptides compared to resultsobtained with the TiO2 resin (Larsen et al., 2005). The IMACloading and rinsing solution consisted of 250 mM acetic acid,30% acetonitrile, and the eluent was NH4OH, pH 10.5. This studyalso compared the use of MALDI-MS and LC-ESI-MS/MS forphosphopeptide analysis. LC-ESI-MS/MS analysis resulted inthe detection of eight phosphopeptides, whereas MALDI-MSdetected>16. Many phosphopeptides that were not detected withLC-ESI-MS/MS were multiply phosphorylated. A similar resultwas also observed in another study (Gruhler et al., 2005).

In addition to examining 2,5-DHB in loading solutions, thecompetitive effects of other acids (some structures are shown inFig. 8) on the binding of nonphosphorylated peptides to TiO2

were also examined (Larsen et al., 2005). These studies gave thefollowing order of ability to inhibit nonphosphopeptide binding:2,5-DHB� salicylic acid� phthalic acid> benzoic acid�cyclohexane carboxylic acid> phosphoric acid>TFA> aceticacid. IR spectroscopy showed that substituted aromatic carbox-ylic acids interact more strongly with TiO2 than do aliphaticcarboxylic acids that contain one –COOH group (Dobson &McQuillan, 2000). Phosphoric acid (Langmuir binding constantof 4� 104 M�1 at pH 2.3) and substituted aromatic carboxylicacids (binding constants of 104–105 M�1) have similar bindingaffinities for TiO2 (Connor & McQuillan, 1999), but thecoordination geometry of salicylate and phosphate to TiO2 aredifferent. Salicylic acid creates a chelating bidentate structurewith the TiO2 surface, whereas a bridging bidentate complex



forms when phosphate (from phosphoric acid) binds to thesurface. Larsen et al. suggested that due to such differences incoordination geometry, 2,5-DHB predominantly competes forbinding sites with nonphosphorylated peptides and not phos-phorylated peptides (Larsen et al., 2005).

Thingholm and co-workers recently provided a protocol toenrich phosphopeptides offline with titania-packed pipette tipsprior to analysis with either LC-MS or MALDI-MS (Thingholmet al., 2006). Some of the main enrichment conditions included aloading solution that contained 100–300 mg/mL of 2,5-DHB in80% acetonitrile, 2% TFA (5% TFA was recommended for morecomplex samples), a washing solution comprised of 80%acetonitrile and 2% TFA, and an elution solution of 0.5%NH4OH (recommended pH should be �10.5). Additionally, theprotocol recommended desalting of an acidified phosphopeptideeluent with POROS Oligo R3, a resin designed for phospho-diester and phosphorothioate oligonucleotide purification(Applied Biosystems, http://www3.appliedbiosystems.com/AB_Home/index.htm, May 6, 2008). For the analysis of enrichedphosphopeptides with MALDI-MS, the suggested matrix was20 mg/mL of 2,5-DHB in 50% acetonitrile, 1% H3PO4. Similarenrichment methods (no offline eluent desalting step wasperformed prior to MS analysis) and analysis with MALDI-MS/MS and nano-LC-ESI-MS/MS were used to identify newphosphorylation sites in proteins isolated from spinach stromamembranes (Rinalducci et al., 2006).

Strong cation exchange (SCX) and titanium dioxidechromatography were utilized for phosphopeptide enrichmentand stable-isotope labeling by amino acids in cell culture(SILAC) for quantitation to study phosphorylation changes uponin vivo stimulation of HeLa cells with epidermal growth factor(Olsen et al., 2006). First, fractions of digested HeLa cell extractswere collected from SCX chromatography, and the phosphopep-tides were further enriched with TiO2-packed tips with 2,5-DHBin the loading solution (rinse solutions consisted of either 10% or80% acetonitrile, 0.1% TFA, and elution solutions containedeither 20% or 40% acetonitrile with NH4OH, pH 10.5). Theenriched fractions were dried, and the residue was reconstitutedin 5% acetonitrile, 0.1% TFA, and subsequently was analyzedwith nano-LC-ESI-MS (multistage MS was used to characterizethe phosphopeptides). This protocol resulted in the detection ofover 10,000 phosphopeptides from over 2,200 proteins. Interest-ingly, this study showed that the relative abundance ofphosphorylation sites of tyrosine, threonine, and serine, basedon more than 2,000 proteins, was 1.8%, 11.8%, and 86.4%,respectively. This result may be a better representation of theabundance of tyrosine phosphorylation than the value of�0.05%found from phosphoamino acid analysis (Hunter & Sefton,

1980). However, we should note that the mass spectrometrystudy was performed with serum-starved cells, which mightchange the relative abundances of the phosphorylated aminoacids.

Although the incorporation of 2,5-DHB in binding and/orrinsing solutions minimizes nonspecific adsorption of acidicnonphosphorylated peptides without hindering phosphopeptidebinding to TiO2 (Larsen et al., 2005), a few recent studiessuggested alternative compounds as ‘‘nonphosphopeptideexcluders’’ that minimize adsorption of nonphosphopeptides(Jensen & Larsen, 2007; Sugiyama et al., 2007; Yu et al., 2007).One of the main reasons to examine new excluders is that 2,5-DHB is not highly compatible with LC-ESI-MS because ofcolumn clogging, loss of sensitivity over time, ion suppression,and coelution of 2,5-DHB with phosphopeptides (Sugiyamaet al., 2007). The incorporation of 300 mg/mL of lactic acid in theloading and washing solution enabled the recovery of 12phosphopeptides from a digest mixture of a-casein, fetuin, andphosvitin (2.5-mg of each protein) with TiO2 enrichment.Importantly, only phosphopeptides were detected. In compar-ison, when 300 mg/mL of 2,5-DHB, rather than lactic acid, wasused, three nonphosphopeptides and only four phosphopeptideswere observed (Sugiyama et al., 2007). When the lactic acidprotocol was applied to a 2.5-mg phosphoprotein digest mixturecontaining a-S1-casein, a-S2-casein, fetuin, and phosvitin, theaverage recovery of 13 phosphopeptides was approximately50%. In experiments with more complex samples, Sugiyamaet al. identified 1100 phosphopeptides from HeLa cells usingenrichment on TiO2 with lactic acid as an excluder in the loadingand rinsing solutions.

When a 3-mg digest of b-casein was applied to a micro-column that contained 5-mm diameter TiO2 particles, and when200 mM ammonium glutamate was incorporated into the rinsingsolution, 84% phosphopeptide recovery was obtained (Yu et al.,2007). This high recovery shows that glutamate does not displacephosphate from the surface, even though it is a diacid. Assuggested earlier, the geometry of some binding sites may bespecific for phosphate groups. A rinsing solution with 130 mM2,5-DHB also afforded high (70%) recovery. The use of theammonium glutamate as an excluder in the analysis of HeLa cellextracts resulted in identification of 858 phosphopeptides, ofwhich 79% were monophosphorylated peptides and theother 21% were multiply phosphorylated (Yu et al., 2007).These percentages agree with a simulation performed with thePhospho.ELM database version 4.0 (Diella et al., 2004). It wasestimated that monophosphopeptides make up roughly 80%of the tryptic phosphorylated peptides of biological samples(Ndassa et al., 2006).

FIGURE 8. Structures of acids used in an attempt to minimize nonspecific binding to TiO2.



Lastly, a series of excluders in the loading/washing solutionswas evaluated for titanium dioxide affinity chromatography priorto MALDI-MS. These studies demonstrated that compared toconventional IMAC, TiO2 is more compatible with a number ofreagents commonly used in the solubilization and digestion ofproteins (Jensen & Larsen, 2007). The sample analyzed was adesalted tryptic digest of 12 proteins (250 fmol of each protein),and excluders examined included phthalic, glycolic, oxalic,lactic, gallic, and citric acids. Among these compounds, 1 Mglycolic acid in the loading/washing solution, which alsocontained 5% TFA and 80% acetonitrile, provided the highestphosphopeptide and lowest nonphosphopeptide recoveries (23phosphopeptides and only two nonphosphopeptides weredetected) with TiO2. Glycolic acid, which is the smallesta-hydroxy carboxylic acid, should form a surface chelate thatis similar to the proposed chelate with 2,5-DHB, and this mayaccount for its effectiveness as an excluder.

Many reagents, such as detergents, surfactants, and salts(with the exception of CHAPS and DMSO) as well as bufferreagents such as EDTA and PBS did not hinder phosphopeptidebinding to the TiO2 (Jensen & Larsen, 2007). Overall, the bestFe(III)–NTA–IMAC recovery of phosphopeptides (18 phos-phopeptides and 1 nonphosphopeptide) was obtained whenthe loading solution contained 0.1% TFA, 50% acetonitrile,and 0.2% RapiGest (surfactant for denaturing protein prior todigestion). However, the number of phosphorylated peptides thatbound to the IMAC resin was reduced (compared to results withRapiGest) in the presence of 1% CHAPS (67% reduction),10 mM EDTA (94% reduction), and PBS (44% reduction).When chromatographic enrichments with ZrO2 and TiO2 werecompared, MALDI-MS signals due to multiply phosphorylatedpeptides were higher and more numerous with TiO2 rather thanZrO2 (Jensen & Larsen, 2007). Monophosphorylated peptidesdid not have a higher affinity for ZrO2 as reported previously (seebelow) (Kweon & Hakansson, 2006).

Because surface properties of metal oxides play acritical role in the binding of phosphopeptides, a few studiesattempted to determine how physical characteristics (particlesize and morphology) influence phosphopeptide adsorption.Nanoparticles might isolate phosphopeptides especiallyefficiently because of their high surface area-to-volumeratio. Crosslinked TiO2 nanoparticles had a twofold higherphosphopeptide binding capacity (�300 mmol of phenylphosphate per gram of crosslinked TiO2 nanoparticles) than5-mm diameter TiO2 particles (Liang et al., 2006). Thesenanoparticles afforded a phosphopeptide recovery of �70%and phosphopeptide MS signals (from casein digests) that were atleast twofold higher than those obtained with the use of largerparticles. The nanoparticle stationary phase was crosslinkedto minimize loss of the particles during the rinsing and elutionsteps.

The above-mentioned studies with titania for phosphopep-tide enrichment were carried out offline, and the sampleswere analyzed with either LC-MS or MALDI-MS. Recently,optimization of conditions for online TiO2-based enrichment offour phosphopeptides from a tryptic digest of the two caseinproteins (1 pmol each) was described (Cantin et al., 2007). Theoverall optimum conditions for these four peptides included aloading buffer consisting of 20% acetonitrile and 2% formic acid,

a wash solution comprised of 80% acetonitrile and 2% formicacid, and an elution solution of 200 mM NH4HCO3. For a morecomplex digest mixture, consisting of an in vitro kinase activatedprotein and its nonactivated form, the optimum bindingbuffer was essentially the same, having an acetonitrile contentof 5–20% and 2% formic acid.

2. Enrichment with Zirconium Dioxide

Like titanium dioxide, zirconium dioxide is positively charged atacidic pH and has a higher binding affinity towards phosphatethan carboxylate anions (Blackwell & Carr, 1991a,b). Theisoelectric points of TiO2 and ZrO2 are �6 and �7, respectively(Vassileva, Proinova, & Hadjiivanov, 1996; Renger et al., 2006).Moreover, ZrO2 has been used previously as a chromatographicresin due to its physical and thermal stability, and thus it is apromising material for phosphopeptide enrichment.

In 2006, microtips that contained zirconium dioxide(Glygen, Columbia, MD) were utilized for the enrichment ofphosphorylated peptides, and the specificity and recoveryachieved was compared to that obtained with Glygen microtipscontaining titanium dioxide (with the same binding, rinsing andelution solutions) (Kweon & Hakansson, 2006). With microtipspacked with 50-mg ZrO2, the optimal procedure for the enrich-ment of 100-pmol tryptic a-casein and b-casein digests prior toanalysis with negative-ion ESI-FT-ICR MS employed a 3.3%formic acid (pH 2.0) binding solution, a water rinse, and a 0.5%piperidine (pH 11.5) elution solution. Under these conditions, tena-casein phosphorylated peptides and only one nonphosphory-lated peptide were detected with the ZrO2 tips, whereas eightphosphopeptides and one nonphosphorylated peptide weredetected with the TiO2 tips. Similarly, when a b-casein digestwas analyzed with ZrO2 tips, five phosphopeptides and twononphosphopeptides were detected, whereas use of TiO2 alloweddetection of only four phosphopeptides and at least onenonphosphopeptide. Overall, the TiO2 microtips were moreselective for enrichment of multiply phosphorylated peptides,whereas the ZrO2 tips enriched primarily monophosphorylatedpeptides. The authors suggested that the selectivity of ZrO2 formonophosphorylated peptides could be due to either the higheracidity of zirconia or the higher coordination number comparedto titania. However, the surface properties of these metal oxidematerials are not well understood.

The enrichments of phosphopeptides from tryptic and glu-Cdigests were also compared with metal oxide resins in pipette tips(Kweon & Hakansson, 2006). As mentioned above, completeglu-C digestion results in only one carboxylate group per peptideand eliminates the problem of unwanted, strong binding byspecies that contain multiple carboxylate groups (Seeley, Riggs,& Regnier, 2005). However, glu-C digestion did not provide anysignificant advantage over tryptic digestion in the enrichment ofcasein phosphopeptides with either ZrO2 or TiO2 tips.

The use of b-hydroxypropanoic acid (b-HPA) as an excluderin the loading solution improved the detection of phosphopep-tides with nano-LC-ESI-MS after enrichment on ZrO2

(Sugiyama et al., 2007). Eleven peptides were detected from a2.5-mg sample of a-casein, fetuin, and phosvitin, whereas signalsfrom nonphosphorylated peptides were not detected. Other



excluders examined in that study included 2,5-DHB, glycolicacid, lactic acid, malic acid, and tartaric acid. Application ofenrichment with ZrO2 and b-HPA to the analysis of proteinsdigested from HeLa cells allowed detection of 1181 phosphory-lated peptides, which is slightly higher than results obtainedwhen titania tips were used (1100 phosphopeptides detected).

In another study of ZrO2-based enrichment, 100 mMdiammonium hydrogen phosphate (pH 9) was used as the elutionsolution for the analysis of casein peptides (digested proteinsincluded a-S1-, a-S2-, b-, and k-casein isolated from milk, andmost of the peptides resulted from nonspecific cleavage due tomilk proteases) (Cuccurullo et al., 2007). After peptides wereeluted, they were acidified and analyzed with either surface-enhanced laser desorption/ionization (SELDI)-MS/MS (normalphase ProteinChip used for sample desalting) or nano-LC-ESI-MS/MS. Four phosphorylated peptides were detected by bothmethods, whereas three additional, unique phosphopeptides weredetected with SELDI-MS, and two with nano-LC-ESI-MS.When SELDI-MS/MS was used 6 individual phosphoserineresidues were detected (theoretically, at least 20 phosphorylationsites among the 4 casein variants are possible), and 5 individualsites were detected with LC-MS/MS.

Nanoparticles of ZrO2 were employed for phosphopeptideenrichment because of their high surface area-to-volume ratio(Zhou et al., 2007). After enrichment with these nanoparticles,the abundance of phosphopeptides was two orders of magnitudegreater than the abundance of nonphosphopeptides, and thismethod can be applied to femtomole levels of protein digests (5–500 fmol of b-casein digest). Application of this technique to theanalysis of a tryptic digest of mouse liver lysate with nano-LC-MS/MS enabled identification of 248 phosphorylation sites from140 phosphopeptides.

3. Enrichment with Aluminum Hydroxide, Alumina,and Niobium (V) Oxide

Compared to studies of TiO2 and ZrO2, relatively little work hasfocused on phosphopeptide enrichment with alumina andAl(OH)3. However, enrichment with Al(OH)3 allowed detectionof eight phosphorylated peptides from an a-casein digest,whereas nonphosphopeptide signals were greatly reducedcompared to conventional MALDI-MS analysis (Wolschin,Wienkoop, & Weckwerth, 2005). More recently, porous anodicalumina membranes were utilized as enrichment materials forphosphopeptides (Wang et al., 2007). Such membranes werepreviously used to desalt biological samples (Wang, Xia, & Guo,2005). Pieces of the membrane were incubated in a peptidemixture for 6 hr and subsequently rinsed with water andacetonitrile for 20 min. After adding the matrix, the membranewas attached to a stainless steel MALDI plate for MS analysis.Although these membranes showed high specificity for mono-phosphorylated (FQpSEEQQQTEDELQDK, m/z 2062, andFQpSEEQQQTEDELQDKIHPF, m/z 2556) and tetraphosphory-lated (RELEELNVPGEIVEpSLpSpSpSEESITR, m/z 3122)peptides from a b-casein digest, the amount of digest requiredto detect all three peptides was 20 pmol. (The peak at m/z 2556was likely due to chymotrypsin cleavage.) A b-casein digest thatcontained 400 fmol was also analyzed, but only produced a single

monophosphopeptide peak (m/z 2062). The authors alsoshowed that 2,5-DHB in 1% phosphoric acid improvedthe detection of the tetraphosphorylated peptide from themembrane relative to the use of an a-cyano-4-hydroxycinnamicacid matrix (a-CHCA); those results are consistent with a priorreport (Wang et al., 2007). Most likely, multiply phosphorylatedpeptides bind more strongly to the alumina membranethan monophosphorylated peptides, and 2,5-DHB and phos-phoric acid are needed to elute these multiply phosphorylatedspecies into the matrix. In conventional MALDI, Kjellstromand Jensen noted that the addition of phosphoric acid to 2,5-DHBmatrix solutions can enhance phosphopeptide signals,including those due to tetraphosphopeptides (Kjellstrom &Jensen, 2004).

A very recent report demonstrates the potential of Nb2O5 asan enrichment medium (Ficarro et al., 2008). With simplesamples containing 1 pmol of standard phosphopeptides,recoveries of phosphopeptides ranged from 50% to 100%. Theuse of lactic acid or 2,5-DHB in loading buffers decreased thebinding of nonphosphopeptides to the resin, and enrichment ofphosphopeptides from a cell lysate revealed several hundredputative phosphorylation sites. Moreover, approximately 30% ofthese phosphorylation sites were not found with enrichment onTiO2, and approximately 30% of the phosphorylation sitesfound with TiO2 enrichment were not identified with Nb2O5.Specific amino acids were associated with sequences of peptidesfound on each of the metal oxides. Thus, the combination ofenrichment on TiO2 and Nb2O5 allows for identification of farmore phosphopeptides than the use of either enrichment materialseparately.

IV. NEW PLATFORMS AIMED TO SIMPLIFYTHE ENRICHMENT PROCEDURE

The previous section provided many examples of the potential ofIMAC and MOAC for phosphopeptide enrichment. However,such procedures frequently require the use of chromatographycolumns and multiple sample-handling steps that complicatethe analysis and present the opportunity for sample losses onthe column walls. In some cases, chromatography might not beappropriate for small sample volumes. One possibility to processsmall samples is the use of pipette tips filled with selective resin,as discussed in the MOAC section above. The subsections belowpresent two alternative platforms, magnetic beads and MALDIplates coated with affinity materials.

A. Magnetic Beads

The use of magnetic beads for phosphopeptide enrichment isattractive because the beads can be easily collected usingan external magnetic field. The development of nano-sizedmagnetic beads makes the technique even more attractivebecause the high surface area-to-volume ratio provides a highbinding capacity. Thus far, beads have been modified with bothmetal affinity complexes and metal oxides. Much of this workwas reviewed very recently (Han, Ye, & Zou, 2008), so we onlypresent a few highlights here.



1. Metal Affinity Beads (IncludingZirconium Phosphonate)

In 2007, Fe(III) and Ga(III) IMAC magnetic beads were preparedusing a tetradentate ligand (not specified), and enrichment withthese modified beads was compared with enrichment by othermaterials, including magnetic beads derivatized with IDA andmetal oxides (Liang et al., 2007). The isobaric tag for relative andabsolute protein quantitation (iTRAQ) labeling technique wasused along with MALDI-MS/MS to determine the recovery ofsimple mixture of phosphopeptides. A recent review describesiTRAQ applications towards phosphoproteomics (Smith &Figeys, 2008). The new Fe(III) IMAC magnetic beads out-performed both similar magnetic beads charged with Ga(III) anda commercial ZrO2 resin (Calbiochem, Gibbstown, NJ), both ofwhich gave phosphopeptide recoveries that were only 10% of thatobtained with the Fe(III) IMAC material. Recoveries achievedwith Fe(III) IMAC magnetic beads were almost identical to thoseof TiO2 materials (Liang et al., 2007), including TiO2 spheresfrom GL Sciences that have been used in other studies (Larsenet al., 2005; Thingholm et al., 2006). Additionally, MALDI-MSsignals due to nonphosphorylated peptides were 50–70% lesswhen enrichment was performed with the new Fe(III) IMACmagnetic beads rather than a pipette tip that contained acommercial IMAC resin with Fe(III)–IDA complexes. Detectionlimits were ca. 5 fmol.

Immobilized metal affinity chromatography (IMAC) andTiO2-coated magnetic beads were also compared by enrichingphosphopeptides from an in-gel digest of tyrosine-immunopre-cipitated (IP) phosphoprotein from cell lysates (Liang et al.,2007). Proteins that contain phosphotyrosine residues weretargeted in this study because they represent <2% of thephosphoamino acids in a cell (Hubbard & Cohen, 1993; Olsenet al., 2006; Zhang & Neubert, 2006). Signals due tophosphopeptides were dominant over signals of nonphosphory-lated peptides, and the MALDI mass spectra looked similar forenrichment by either the new IMAC or TiO2-coated beads.Additionally, this study identified six new phosphorylation sitesupon epidermal growth factor stimulation.

Fe(III)–IDA-coated magnetic beads were also synthesizedwith particle diameters of only 15 nm to increase the surface area-to-volume ratio�600-fold relative to micron-sized particles (Tanet al., 2008). In MALDI-MS analyses of 2 pmol, 1 pmol,100 fmol, and 20 fmol of a-casein digest, the Fe(III)–IDAmagnetic nanoparticles enabled detection of 17, 13, 7, and2 phosphopeptides, respectively. In contrast, at one pmol ofa-casein digest, the use of the commercial IMAC resin andMALDI-MS detected 5 phosphopeptides, whereas at 100 fmol,no phosphopeptides were detected. Additionally, when 2 pmol ofa-casein digest was analyzed in the presence of 20 pmol of BSAdigest, the magnetic beads facilitated detection of 14 phospho-peptides, whereas the commercial IMAC resin revealed half thatnumber. The enrichment protocols used for the two materialswere essentially identical; however, the loss of commercial resinduring the enrichment steps (rinsing and elution) might haveoccurred because no mechanism was used to securely retain theparticles while the supernatant was drawn off from the beadslurry. The lowest amount of the synthetic monophosphopeptidethat could be analyzed with the magnetic Fe(III)–IDA nano-

particles and MALDI-MS was 5 fmol (Tan et al., 2008). After adigest of plasma membrane proteins from mouse liver wasfractionated with SCX chromatography and desalted with a C18

cartridge, the samples were analyzed with the IMAC magneticnanoparticles and nano-LC-MS/MS. This analysis resulted in theidentification of 217 phosphorylation sites from 158 phospho-proteins.

The synthesis and use of zirconium phosphonate-modifiedsubstrates (monolithic capillary columns, polymer beads,MALDI plates, and magnetic beads) as new IMAC materialsfor phosphopeptide enrichment and subsequent MS analysis havebeen reported in the last few years (Zhou et al., 2006; Dong et al.,2007; Feng et al., 2007; Wei et al., 2008). The most recent ofthese publications involved the use of magnetic nanoparticles(�70-nm particle diameter) modified with Zr(IV)-phosphonatecomplexes to enrich femtomole levels (50–500 fmol) of b-caseinphosphopeptides (Wei et al., 2008). The detection limit with theZr(IV)-phosphonate magnetic beads and MALDI-TOF-MS was�50 fmol based on the detection of a single monophosphorylatedpeptide (m/z 2062) from a b-casein digest. However, in thepresence of a 100-fold excess of BSA digest, signals due to anumber of nonphosphorylated peptides were observed in themass spectrum when phosphopeptides were enriched from500 fmol of b-casein digest. With iTRAQ reagents, the capacityof the nanoparticles and the phosphopeptide recovery ofa standard protein, FLpTEYVATR, were determined to be141 pmol/mg and �53%, respectively. Lastly, to show theapplicability of this enrichment method towards large-scalebiological samples, a SCX chromatographic fraction from digest-ed proteins from a Chang liver cell extract was enriched with theZr(IV)-phosphonate magnetic nanoparticles and analyzed usingnano-LC-ESI-MS. This method detected 22 phosphorylatedpeptides from one SCX fraction (buffer contained 1 M NaCl).Enrichment with commercial PHOS-select Fe(III)-IMAC beadsfrom Sigma allowed detection of only six phosphopeptides.

2. Metal Oxide-Coated Beads

The use of magnetic nanoparticles coated with metal oxides(e.g., titania, zirconia, alumina, and gallium oxide) to isolatephosphopeptides began in 2005 (Chen & Chen, 2005). As shownin Figure 9, magnetic (Fe3O4) nanoparticles were coated with athin layer of SiO2 via either TEOS or sodium silicate, and themetal oxide, TiO2, ZrO2, Al2O3, or Ga2O3, was then formed onthe silica with titanium butoxide, zirconium butoxide, aluminumisopropoxide, or gallium isopropoxide, respectively (Chen &Chen, 2005, 2008; Chen et al., 2007; Lo et al., 2007; Li et al.,2008). This procedure typically formed �50-, �170-, and �20-nm diameter titania-, zirconia-, and alumina-coated Fe3O4

particles, respectively. Similar ZrO2- and Al2O3-coated magneticnanoparticles with diameters of �280 nm have also beenprepared (Li et al., 2007a,b).

The protocol to use the metal oxide-coated magneticnanoparticles to enrich phosphopeptides is fairly simple. Forexample, 24 mg of magnetic beads was mixed with �50 mL ofprotein digest in 0.15% TFA in a microcentrifuge tube, incubatedfor 30 sec, rinsed with 50% acetonitrile that contained 0.15%TFA, and the solution was decanted while using a magnet



to secure the nanoparticles to the wall of the tube (Chen et al.,2007). The beads were either applied to a MALDI plate withoutthe addition of matrix (Fe3O4/TiO2 surface-assisted laserdesorption/ionization, SALDI) or they were mixed with 2,5-DHB that contained phosphoric acid and applied to the MALDIplate for direct MALDI-TOF-MS analysis (Chen & Chen, 2005).The Fe3O4/TiO2 SALDI method is attractive because no elutionstep or matrix is necessary.

The estimated binding capacity of alumina-coated magneticnanoparticles is 60 mg of phosphopeptide per milligram ofnanoparticle (Chen et al., 2007). Using this capacity, 25 mg ofmagnetic beads could isolate 1.5mg of phosphopeptide, which fora phosphopeptide with a molecular weight of 2500 Da,corresponds to 600 pmol. This high capacity could lead tononspecific binding. The binding capacities for the other metaloxide-coated magnetic nanoparticles were not specified.

Most studies of magnetic beads employed relatively largeamounts of simple phosphopeptide mixtures such as a- and b-casein digests. Recently, magnetic nanoparticles coated withTiO2 and Al2O3 were used to enrich phosphopeptides fromthree batches of diluted human serum with only a 30-secincubation time (Chen & Chen, 2008). Although the differentspecificities of the two metal-oxide coated magnetic materialswere apparent from variations in the relative peak intensitiesin the mass spectra, both particles enriched four phosphorylatedvariations of fibrinopeptide A (DpSGEGDFLAEGGGV atm/z 1390, ADpSGEGDFLAEGGGV at m/z 1460, DpSGEGD-FLAEGGGVR at m/z 1546, and ADpSGEGDFLAEGGGVR atm/z 1617). Additionally, a peak at m/z 2753 was present inthe analysis of the three batches of blood serum only with theTiO2-coated magnetic nanoparticles. This peak was due to anonphosphorylated human serum albumin peptide (DAHKSE-VAHRFKDLGEENFKALVL). However, because no 2,5-DHBwas used in the loading or rinsing solutions, the nonspecificadsorption observed with the TiO2 magnetic nanoparticles mightbe reduced if the appropriate excluder were included in thesesolutions.

B. On-Plate Enrichment for MALDI-MS

This section presents recent developments in the on-plate affinitycapture of phosphopeptides for subsequent MALDI-MS analy-sis. On-plate enrichment is advantageous because it allows

for minimal sample handling, high throughput, and lower sampleloss than conventional resin-based techniques. With a conven-tional IMAC technique, 10–15% of phosphopeptides were lostduring the rinsing step, an additional 10–20% were still retainedon the IMAC column after elution, and another 10–20% of thephosphopeptides were lost when samples were desalted with C18material (Kokubu et al., 2005). The on-plate enrichment processshould result in less sample loss because it simply consists ofincubation on the plate, rinsing of the sample on the plate, andaddition of matrix. The initial modification of MALDI samplesupports as affinity devices for the isolation of biologicalmolecules with subsequent analysis by MS began in the early1990s and was carried out by a number of groups (Hutchens &Yip, 1993; Papac, Hoyes, & Tomer, 1994; Brockman & Orlando,1995, 1996; Dogruel, Williams, & Nelson, 1995; Nelson et al.,1995). Subsections below describe isolation of phosphopeptideson MALDI plates modified with both IMAC functionalities andmetal oxides.

1. IMAC MALDI Plates

a. Commercial MALDI affinity chipsCiphergen Biosystems (Palo Alto, CA) developed a ProteinChipSystem that consists of a ProteinChip Reader (i.e., MALDI linearTOF mass spectrometer), ProteinChip Software, and relatedaccessories that are used in conjunction with ProteinChip Arrays(Tang, Tornatore, & Weinberger, 2004). A number of reportsdescribed the use of these chips for phosphorylated peptide andprotein purification (Cardone et al., 1998; Davies, 2000; Laineet al., 2000; Roig et al., 2000; Mortier et al., 2001; Stoica et al.,2001; Tassi et al., 2001; Espina et al., 2004; Liu et al., 2004;Thulasiraman et al., 2004; Bowley et al., 2005; Ge, Gibbs, &Masse, 2005; Head et al., 2005; Righetti et al., 2005; Wu, Jin, &Marsh, 2005; Le Bihan et al., 2006; Bodega et al., 2007; Akashi &Yamori, 2007; Akashi et al., 2007). One example was the use ofan IMAC-Gallium ProteinChip Array (IMAC chips were chargedwith a solution of 50 mM gallium nitrate) for the simultaneousanalysis of a mixture of three synthetic monophosphorylatedpeptides, KRPpSQRHGSKY, TRDIYETDYpYRK, and KREL-VEPLpTPSGEAPNQALLR, and their nonphosphorylated coun-terparts (Thulasiraman et al., 2004). The nonphosphorylatedpeptides (1–3 pmol) were in a 10-fold excess compared to thephosphorylated analogs, but all three phosphorylated peptides

FIGURE 9. Fabrication of TiO2-coated magnetic nanoparticles (procedure from Chen & Chen, 2005,

2008). Zirconium, aluminum, and gallium oxide-coated magnetic nanoparticles were prepared in a similar

fashion (Chen et al., 2007; Li et al., 2007a,b, 2008; Lo et al., 2007; Chen & Chen, 2008).



showed greater signals in the mass spectrum than theirnonphosphorylated counterparts. Additionally, peptide sub-strates phosphorylated with three specific kinases could bedetected simultaneously with the IMAC ProteinChip Array. Theuse of these chips requires the Ciphergen SELDI ion source,which can be coupled to other commercial mass spectrometers.These chips and equipment are currently available through Bio-Rad (Hercules, CA).

Qiagen (Valencia, CA) offers phosphopeptide-affinity platesthat are compatible with several MALDI mass spectrometers.The sample wells of the Qiagen IMAC Mass Spec Focus Chipscontain a hydrophobic zone that encloses up to 35 mL of samplewithin the well along with an affinity zone that contains ligandsthat, when charged with a solution of Fe(III), capture thephosphopeptides. When the bound phosphopeptides are elutedfrom the affinity zone with matrix, they are concentrated orfocused into a 0.6-mm diameter analysis zone, which is morehydrophilic than the affinity zone. Although the affinity ligandson the chip are not described, in 2006 Qiagen presented a poster atthe American Society for Mass Spectrometry meeting on the useof chips modified with NTA self-assembled monolayers (SAMs)for phosphopeptide purification and concentration (Belisle et al.,2006).

We recently applied these Qiagen IMAC chips towards therecovery of 125 fmol of a synthetic phosphopeptide from a1 pmol protein-digest mixture (BSA, phosphorylase b, andesterase) (Dunn et al., 2008). Using an isotopic internal standard,the recovery of the phosphopeptide from a protein digest mixturewas 13%, and minimal signals due to nonphosphorylatedpeptides were observed. More recent results from a digestcontaining threefold higher reagent concentrations showed arecovery of 30%. We estimate that the binding capacity of theQiagen IMAC chip is ca. 5 pmol. This estimate assumes anaffinity zone area of 7 mm2 and binding of a monolayer ofphosphopeptide to this zone with a density of 1 peptide per2.5 nm2. Although the Qiagen IMAC chip has excessbinding capacity for a 125 fmol sample, recovery was still 30%or less.

b. Monolayers of metal-ion complexes immobilized onMALDI surfacesA number of research groups have prepared custom MALDIplates with IMAC functionalities. The immobilization of SAMsderivatized with NTA analogs was first reported in 1996, and anumber of subsequent studies demonstrated the fabrication ofsimilar monolayer films (Sigal et al., 1996; Stora et al., 1997;Scheibler et al., 1998; Kada et al., 2002; Luk et al., 2003;Makower et al., 2003; Rigler et al., 2003; Wegner et al., 2003;Gamsjaeger et al., 2004; Lee et al., 2004; Rigler, Ulrich, & Vogel,2004; Trammell et al., 2004; Maly et al., 2004a,b; Johnson &Martin, 2005; Tinazli et al., 2005; Klenkar et al., 2006; Valiokaset al., 2006; Ataka & Heberle, 2006; Ataka, Richter, & Heberle,2006). These NTA SAMs have been employed in the study ofhistidine–nickel, protein–protein, protein–antibody, protein–DNA, or protein– ligand interactions, and for biotechnologyapplications including microarrays, biosensors, catalysis, andbiocompatible coatings. However, it was not until recently thatNTA SAMs immobilized on gold MALDI plates were used tocapture phosphorylated peptides for direct MALDI-TOF-MSanalysis (Shen et al., 2005). In short, a SAM of 1,8-octanedithiolwas deposited on a gold substrate, and allowed to react withmaleimide-terminated NTA (N-[5-(30-maleimidopropylamido)-1-carboxypentyl]-iminodiacetic acid) as shown in Figure 10. Theimmobilized NTA ligand was subsequently charged with Ga(III)from 200 mM gallium nitrate. Shen and co-workers used theseGa(III)–NTA–SAM-modified plates for on-probe enrichmentof a mixture that contained two synthetic phosphopeptides,DLDVPIPGRFDRRVpSVAAE and KIGDFGMTRDIYETDpY-pYRKGGK, and four nonphosphorylated peptides, angiotensin I,ACTH 1–17, ACTH 18–39, and ACTH 7–39 (ACTH isan abbreviation for adrenocorticotropic hormone). In theconventional MALDI-MS analysis of the mixture, the fournonphosphorylated peptides gave stronger signals than themonophosphorylated peptide, and the diphosphorylatedpeptide was not detected. In contrast, both phosphopeptideswere detected when the same mixture was analyzed after captureon the Ga(III)–NTA–SAM-modified probe, whereas the peaks

FIGURE 10. Formation of a self-assembled monolayer of octanedithiol and derivatization of this

monolayer by reaction with maleimide-terminated NTA (procedure from Shen et al., 2005). Further NTA

complexation of Ga(III) or Fe(III) is not shown here.



due to the nonphosphorylated peptides were eliminated orreduced. (The matrix employed in their analyses was a-CHCA.)However, even though the MS signal due to the nonphosphory-lated peptide ACTH 18–39 was significantly reduced, it was stillmore intense than the peaks due to the phosphorylated peptides.

A tryptic digest of b-casein was also analyzed with theGa(III)–NTA–SAM-modified plates (Shen et al., 2005). Con-ventional MALDI-MS analysis generated signals for a fewnonphosphorylated peptides in addition to a relatively low-intensity peak for the monophosphorylated peptide. The signalfor the monophosphorylated peptide increased approximatelythreefold when the SAM-modified probe was used ratherconventional MALDI-MS analysis. However, no signal forthe tetraphosphorylated peptide was observed in either theconventional analysis or analysis after enrichment on themodified plate. Nonspecific adsorption of nonphosphorylatedpeptides was also apparent in the mass spectrum when the digestwas applied to the Ga(III)–NTA–SAM-modified probe.The authors stated that they saw better reproducibility in themass spectra when Ga(III) rather than Fe(III) was used as themetal ion.

In a completely different synthetic method, Xu and co-workers derivatized a porous silicon surface with a monolayer ofIDA-1,2-epoxy-9-decene via a photochemical reaction (Xu et al.,2006). In brief, excess IDA was allowed to react with 1,2-epoxy-9-decene to form a photochemically reactive IDA derivative. Theelectrochemically etched porous silicon substrate was immersedin a solution of the IDA derivative, and was exposed to light froma 1000-W Hg lamp for 2 hr, after which IDA-1,2-epoxy-9-decenederivative was assumed to be immobilized onto the siliconsurface as shown in Figure 11. The IDA-derivatized siliconsurface was immersed in a 100 mM FeCl3 solution to form theFe(III)–IDA complex.

The Fe(III)–IDA-derivatized porous silicon plates wereused to analyze tryptic digests of b-casein with a matrix solutionthat contained 2,5-DHB and 1% phosphoric acid. The conven-tional mass spectrum of the digest revealed the presence of thethree phosphorylated peptides with m/z 2062, 2556, and 3122along with several peaks due to nonphosphorylated peptides.When a one pmol digest was analyzed after capture on theFe(III)–IDA-modified silicon probe, only peaks due to thephosphorylated peptides were present. However, there wassignificant background noise in the mass spectrum, and thesignal intensity due to the phosphopeptides decreased comparedto the conventional mass spectrum. When the amount of digestwas decreased from 1 pmol to 300 fmol, all three phosphopep-tides were still detected when capture on the modified siliconplate was used, and nonspecific adsorption from other peptideswas minimal. At the lower amount of digest, the peak intensities

of the phosphopeptides were higher than those observed in theconventional MALDI mass spectrum of the same amount ofdigest.

c. Polymer-modified MALDI platesTo increase the binding capacity for metal-affinity-based on-plateenrichment, we recently modified Au plates with polymer filmsthat contain metal-ion complexes. Initially, we used patternedplates with small (0.2-mm diameter) spots of a hydrophilicpolymer, poly(acrylic acid) (PAA), surrounded by a hexadeca-nethiol monolayer (Xu, Bruening, & Watson, 2004). Immersionof the plate in an Fe(NO3)3 solution created the Fe(III)–PAAcomplex. These patterned metal-affinity probes served to bothpurify and concentrate the analyte as shown in Figure 12 (Xu,Watson, & Bruening, 2003; Xu, Bruening, & Watson, 2004).

When the patterned, Fe(III)–PAA-modified plate was usedto enrich the phosphopeptides from one pmol of ovalbumindigest, signals due to the phosphopeptides were enhancedcompared to those observed in the conventional MS analysis;however the mass spectrum of the enriched sample still containedlarge signals due to nonphosphorylated peptides (Xu, Bruening,& Watson, 2004). The matrix used in these phosphopeptideanalyses was a-CHCA. The surface might not be completelysaturated with Fe(III), and excess negatively charged carboxylategroups might bind positively charged, nonphosphorylatedpeptides via electrostatic interactions. Perhaps more stringentrinsing, including acetic acid and/or acetonitrile, could havealleviated nonspecific adsorption.

Poly(acrylic acid) (PAA)-modified plates derivatized withpoly(ethyleneimine) (PEI) were also used to capture phosphory-lated peptides from 100 fmol of b-casein digest (Xu, Bruening, &Watson, 2004). The signals due to the monophosphorylated andtetraphosphorylated peptides of b-casein dominated the massspectrum of a digest captured on the PEI-modified plate. In thiscase, enrichment likely occurred because of interactions betweenpositively charged PEI and the negatively charged phosphopep-tides. In addition to the phosphate groups, there are several acidicamino acid residues (D and E) in the tryptic phosphopeptidesof b-casein, and these groups could also be attracted to thepositively charged PEI surface. It is doubtful that an electrostatictechnique could be used to purify typical phosphopeptides withfewer acidic sites in moderately complex mixtures.

To enhance the selectivity of polymer-modified plates, wederivatized thin PAA films (�30 A) with the Fe(III)–NTAcomplexes that are frequently employed in IMAC (Dunn,Watson, & Bruening, 2006). Deposition of protein digests onthe polymer-modified plates, followed by a rinse with an aceticacid solution, addition of matrix, and subsequent analysis withMALDI-MS produced mass spectra that were dominated

FIGURE 11. Immobilization of a monolayer of an IDA derivative on porous silicon using photochemistry

(procedure from Xu et al., 2006).



by peaks due to singly and multiply phosphorylated peptidesfrom b-casein and ovalbumin digests.

We have also examined on-plate enrichment with poly(2-hydroxyethyl methacrylate) (PHEMA) brushes modified withFe(III)–NTA complexes (Dunn et al., 2008). These brushes are10-fold thicker than the Fe(III)–NTA–PAA films mentionedabove, and thus have greater binding capacities. Ellipsometricstudies of phosphoangiotensin binding to these films revealed abinding capacity of �0.6 mg/cm2, and enrichment with theFe(III)–PHEMA–NTA brushes allowed MALDI-MS detectiondown to 15 fmol of phosphopeptide (based on the detection of amonophosphopeptide (m/z 2062) from a b-casein digest). Thehigh sensitivity is presumably due to a relatively high bindingcapacity. Using an isotopically labeled internal standard added tothe MALDI plates after rinsing and prior to addition of matrix, wedetermined a recovery of �70% for a synthetic monophospho-peptide with the Fe(III)–NTA–PHEMA-modified plates, evenwhen the digest contained one pmol of BSA, phosphorylase b,and esterase. Unfortunately, recovery does decrease greatly in thepresence of higher amounts of digest reagents. In contrast, thinfilms of Fe(III)–NTA–PAA recovered just �23% of thesynthetic phosphopeptide, and a monolayer of Fe(III)–NTArecovered only �9%. We also compared the performance of theFe(III)–NTA–PHEMA-modified plates with commerciallyavailable IMAC and metal oxide materials. The phosphopeptiderecoveries of the commercial enrichment methods were typicallythreefold lower than that of the PHEMA brushes, with theexception of a TiO2 microtip, which had a similar recovery of�70%. However, the TiO2 microtip exhibited significantimpurities in its eluent. Lastly, the analysis of a synthetic

diphosphorylated peptide resulted in �100% recovery from aprotein digest mixture that contained BSA, phosphorylase b, andesterase when enrichment was performed using the Fe(III)–NTA–PHEMA-modified plates. Because they have multiplebinding sites, diphosphorylated peptides evidently have a higheraffinity for the Fe(III)–NTA complex than do monophosphory-lated peptides.

Recently, a plastic MALDI chip that selectivelycaptures phosphorylated species was fabricated (Ibanez, Muck,& Svatos, 2007). The chips were prepared by copolymerizingmethyl methacrylate and acrylic acid N-hydroxysuccinimideester, which can react with amino-terminated NTA to providethe immobilized ligand (Ibanez, Muck, & Svatos, 2007). Anyremaining active ester sites reacted with tris(hydroxymethyl)a-minomethane, and the NTAwas charged with Ga(III) or Ni(II) toanalyze phosphorylated peptides or proteins, respectively.Chips charged with Ga(III) were selective for phosphopeptidesfrom an a-casein digest, whereas Ni(II) was better forphosphoprotein recovery based on the enrichment of five pmolof a-casein from an equimolar mixture that contained myoglobinand carbonic anhydrase I (phosphoprotein adsorption occurred atpH �8). The Ga(III)–NTA plastic MALDI chips were used toanalyze one pmol of phosphorylated angiotensin from anonphosphorylated BSA digest (3.5 pmol). After the samplewas applied to the modified plate, rinsed to remove unboundspecies, and mixed with matrix, the dominant signal in themass spectrum was that due to phosphoangiotensin,whereas many signals due to BSA were reduced compared toconventional MALDI-MS. This technique was also used toanalyze �2 pmol of an a-casein digest. Although the plate was

FIGURE 12. Fabrication of polymer-modified gold MALDI plates for phosphopeptide capture and

concentration (Xu, Watson, & Bruening, 2003). A hydrophobic pattern is created first, followed by the

immobilization of polymers onto the 0.2-mm diameter gold wells. HDT¼ hexadecanethiol, PDMS¼ po-

poly(dimethylsiloxane). Reproduced from Xu, Watson, & Bruening, 2003. Copyright 2003, American

Chemical Society, used by permission.



able to retain eight phosphopeptides, some of the strongest peaksin the mass spectrum were due to nonphosphorylated peptides.Thus, the polymer plate appears to be subject to nonspecificadsorption.

d. Zirconium phosphonate filmsA number of studies demonstrated the formation of zirconiumphosphonate films on various surfaces because of potentialapplications in catalysis, sensing, electronics, protein immobi-lization, and separations (Lee et al., 1988; Putvinski et al., 1990;Hong, Sackett, & Mallouk, 1991; Byrd, Pike, & Talham, 1993;Frey, Hanken, & Corn, 1993; Byrd et al., 1994; Nixon et al., 1999;Clearfield et al., 2000; Kohli & Blanchard, 2000; Kumar &Chaudhari, 2000; Benitez et al., 2002; Nonglaton et al., 2004;Mazur, Krysinski, & Blanchard, 2005). Zr(IV) binds stronglyto phosphonate monolayers due to metal ion-ligandcrosslinking (Zr(IV) ions coordinate to more than one phospho-nate molecule), (Byrd, Pike, & Talham, 1993; Byrd et al., 1994)and multilayer Zr(IV) phosphonate films can also be prepared(Nixon et al., 1999; Clearfield et al., 2000; Kohli & Blanchard,2000). With regard to phosphopeptide enrichment, Zhou and co-workers used zirconium phosphonate monolayers immobilizedon porous silicon as phosphopeptide affinity probes for MALDI-TOF-MS (Zhou et al., 2006). To form the phosphonate-siliconsurface, an etched silicon substrate was first placed in a solutionof phosphorous oxychloride (POCl3) and 2,4,6-trimethylpyri-dine (collidine). Subsequently, the surface was charged withZr(IV) by immersing the phosphonate-modified silicon substrateinto 20 mM ZrOCl2, and these modified plates were used toanalyze digests of b-casein with 2,5-DHB in 1% phosphoric acidas a matrix. Enrichment of phosphopeptides from the b-caseindigest with the Zr(IV)-phosphonate silicon substrate yieldedsignals for mono (m/z 2062, 2556) and tetraphosphorylated (m/z3122) phosphopeptides at the 2-pmol and 20-fmol levels. When2 fmol of b-casein was analyzed using the porous silicon plates,only phosphopeptide peaks at m/z 2061 and 2556 were present.Signals due to nonphosphorylated peptides were minimal at alldigest concentrations.

Fifteen phosphorylated peptides were isolated from a 2-pmol a-casein digest with the Zr(IV)-phosphonate-modifiedsilicon plate (Zhou et al., 2006). A number of these peptides arisefrom missed cleavages. Even though 14 of the phosphorylatedpeptides were observed in the conventional MALDI massspectrum, the modified silicon plate simplified the mass spectrumby eliminating nearly all signals due to nonphosphorylatedpeptides.

Additionally, b-casein was combined with a tryptic digest ofBSA, a nonphosphorylated protein, and analyzed at b-casein toBSA molar ratios of 1:1, 1:10, and 1:100 (the amount of b-caseinin the digest was maintained at 1 pmol). Impressively, the use ofthe derivatized porous silicon plates eliminated nearly all signalsthat corresponded to peptides from BSA. However, when theamount of BSAwas 100-fold greater thanb-casein, the signals forthe monophosphorylated peptides (m/z 2061 and 2556) dramat-ically decreased. For comparison, they also applied the mixture toFe(III)-IMAC beads (Poros MC beads, Applied Biosystems,Foster City, CA) that contain an IDA ligand. The Fe(III)–IDAbeads suffered from nonspecific adsorption at high amounts of

the BSA digest (10- and 100-pmol amounts). Nonspecificadsorption is a frequent limitation of IMAC, but the extent ofnonspecific adsorption likely depends on the rinsing and loadingprotocol.

2. MALDI Plates Modified with Metal Oxides

a. Zirconium oxide-functionalized MALDI surfacesRecently, zirconium oxide was immobilized on a stainless steelplate via collisions of Zr(IV)-n-propoxide cations from an ESIsource onto an oxidized surface (Blacken et al., 2007). Thisprocedure resulted in zirconium oxide spots with a diameter of5.5 mm, and these spots showed affinity toward phosphopeptidesin a 1 pmol a-casein digest. The digest was applied to themodified surface, incubated for 20 min, rinsed with an aqueoussolution of 0.1% TFA in 20% acetonitrile, and dried. Addition ofmatrix and MALDI-MS resulted in a mass spectrum with greatlyreduced signals (>25-fold reduction compared to conventionalMALDI-MS) due to nonphosphopeptides, while two mono-phosphorylated peptides (m/z 1660 and 1951) produced two ofthe most intense peaks in the mass spectrum. However, theoverall signal intensity of these two phosphopeptides decreasedby a factor of two compared to peaks in the conventional analysisof the digest. Although this functionalized surface retainedphosphorylated peptides, several synthetic peptides that con-tained methionine or tryptophan residues were readily oxidized,perhaps because of oxidation catalysis by zirconia. The authorssuggested the addition of a reducing agent such as dithiothreitolto the loading solution to minimize any amino acid oxidation.

b. Immobilized titanium dioxide on MALDI platesAs described in Section II.C.1., TiO2 is very effective inphosphopeptide enrichment. In an effort to simply phosphopep-tide isolation, TiO2-coated gold nanoparticles immobilized on aglass slide were employed for on-plate enrichment and analysisof phosphopeptides with MALDI-TOF-MS (Lin, Chen, & Chen,2006). The scheme in Figure 13 shows the fabrication of theTiO2-gold-nanoparticle (TiO2-gold-NP) glass plate. Using theFrens method (Frens, 1973), gold nanoparticles were preparedwith an average diameter of 26 nm. To prepare the glass slide forthe attachment of the gold nanoparticles, a thin film of TMSPED(N-[3-(trimethoxysilyl)propyl]ethylenediamine) was bound tothe oxidized glass surface. Gold nanoparticles were immobilizedonto the TMSPED-treated surface, and a solution of titaniumisopropoxide was spin-coated onto the substrate, followed byannealing. The coverage of the 26-nm diameter gold nano-particles on the glass slide was estimated to be �620 nano-particles/mm2 (34% coverage) (Lin, Chen, & Chen, 2006).Tryptic digests were applied to the modified glass plates in 500-mL aliquots, and after the sample had incubated for 1 hr, thesurface was rinsed to remove any unwanted species prior tothe addition of 2,5-DHB that contained H3PO4 and MALDI-TOF-MS.

The TiO2-gold-NP-modified glass plates were used toanalyze b-casein digests that contained relatively highamounts of protein (50 and 500 pmol) (Lin, Chen, & Chen,2006). Three phosphorylated peptides (m/z 3122, 2556, and2062) were captured and detected with the TiO2-gold-NP plate,



but nonspecific adsorption was observed with the 500-pmolsample (500 mL of 1-mM b-casein digest). Using the modifiedglass plate, Lin and co-workers also analyzed two more complexsamples: (1) an equimolar mixture (50 pmol of each) ofcytochrome c and b-casein, and (2) milk, which contains a-S1-,a-S2-, and b-casein. Interestingly, there were almost no peaksdue to peptides from cytochrome c (a nonphosphorylated protein)in the mass spectrum when the equimolar digest mixture wasenriched with the TiO2-gold-NP-modified glass plate. When themilk digest was applied to the modified plate, four phosphopep-tides (m/z 3122, 3008, 2556, and 1952), one due to chymotrypsincleavage and one due to miscleavage of a-S1-casein, weredetected. This number of phosphopeptides detected is quite lowbecause a-S1- and a-S2-casein are highly phosphorylated (thereare over 20 phosphorylation sites between the two proteins). If allthree casein proteins are digested completely (i.e., no miscleav-age), then 11 phosphorylated peptides should result. Hence, only18% of fully cleaved, phosphorylated peptides were detected inthe milk digest enriched on the titania bead-modified target.Conventional MALDI-MS analyses of the digests would beuseful for comparison.

More recently, an array of titanium dioxide nanoparticleswas formed on a commercial stainless steel MALDI plate

by spotting an array of 2-mL droplets of a 100-mg/mL TiO2

suspension onto the surface, and heating at 4008C for 1 hr (Qiaoet al., 2007). The selectivity of this metal oxide affinity plate wasdemonstrated by enriching phosphopeptides from b-caseindigests at picomole down to low femtomole levels in just30 min (sample incubation time). At relatively high digestconcentrations (850 nM), six phosphopeptides from a-caseinimpurities (m/z 1467, 1540, 1595, 1661, 1928, and 1952) inaddition to three b-casein phosphopeptides (m/z 2062, 2556, and3122) were enriched from a b-casein digest. For a lowerincubation time of 5 min, a single monophosphorylated peptidewas still detected at 30 fmol of b-casein digest. Mostimpressively, the three b-casein phosphopeptides at 26 nM werestill enriched in the presence of a 100-fold excess of BSA digest(30-min incubation).

Polymeric MALDI plates that contained channels packedwith titanium dioxide were used to demonstrate the feasibility ofthis modified plate geometry to enrich phosphopeptides (Ekstromet al., 2007). Similar plates for solid-phase microextraction wereused previously by the same group (Ekstrom et al., 2004, 2006).Plates fabricated with pyramidal-shaped channels were packedwith titania (Fig. 14) to selectively enrich and concentratephosphopeptides from one pmol and 100 fmol b-casein digests

FIGURE 14. Cross-section of a conducting polymer MALDI plate, showing one pyramidal-shaped

channel packed with 50-mm diameter TiO2 particles (Ekstrom et al., 2007). The inlet and outlet have widths

of 1 mm and 50 mm, respectively. All solutions are pulled through the channel using a vacuum.

FIGURE 13. Immobilization of titania-coated gold nanoparticles (NPs) on glass substrates (procedure

from Lin, Chen, & Chen, 2006).



(Ekstrom et al., 2007). A vacuum was used to pull all solutions(digest, wash, elution, and matrix) through the packed channels.Hence, the matrix crystallized on the rear side of the modifiedplate (Fig. 14), which was then analyzed with MALDI-MS.When one pmol of b-casein digest was analyzed, the mono-phosphopeptide (m/z 2061) and tetraphosphopeptide (m/z 3122)were retained and detected, whereas essentially all signals due tononphosphopeptides were eliminated. When the sample amountwas reduced to 100 fmol, only a weak signal due to themonophosphorylated peptide was detected.

TiO2-coated magnetic nanoparticles have also been affixedto the sample wells of a conventional stainless steel MALDI plateby holding a magnet to the back of the sample plate during thesample loading, incubation (10 min), rinsing, and matrix-loadingsteps (Tan et al., 2007). These modified plates were used toenrich phosphopeptides from a-casein, b-casein, and ovalbumindigests. The sensitivity of this method was demonstrated by theenrichment of both the monophosphorylated (m/z 2062)and tetraphosphorylated peptides (m/z 3122) from 100 fmol ofb-casein digest, and the detection limit of the monophosphory-lated peptide appears to be �10 fmol. The selectivity wasdemonstrated by the detection of 14 phosphopeptides, 6 ofwhich were multiply phosphorylated, from a mixture of a-casein,b-casein, ovalbumin, myoglobin and BSA digests. ConventionalMALDI-MS yielded signals for only five monophosphorylatedpeptides. Although the intensities of many signals due tononphosphopeptides were reduced compared to conventionalanalysis, a few nonphosphopeptide peaks gave more intensesignals than those of the phosphorylated peptides.

3. MALDI Matrices

The selection of the MALDI matrix affects phosphopeptidesignals in MALDI regardless of whether on-plate purification isemployed. Typical matrix molecules include a-CHCA, 2,5-DHB, and 20,40,60-trihydroxyacetophenone (THAP). Yang andco-workers showed that the signals of phosphopeptides inb-casein and protein kinase C (PKC)-treated mouse cardiactroponin I increased 10-fold when using THAP with ammoniumcitrate rather than a-CHCA as a matrix (Yang et al., 2004). Wealso found that THAP gives higher phosphopeptide signals thana-CHCA (Dunn, Watson, & Bruening, 2006). Hart and co-workers examined the use of both 2,5-DHB and a-CHCA to elutephosphopeptides from IMAC beads and serve as a matrix. In theanalysis of an a-casein digest, DHB revealed more peptides withstronger signals. They suggested that 2,5-DHB gives reducedenergy transfer to the matrix to yield more intact phosphopep-tides (Hart et al., 2002).

Kjellstrom and Jensen demonstrated the use of phosphoricacid, after examining five acids (acetic acid, formic acid, TFA,heptafluorobutyric acid, and phosphoric acid), as a suitablematrix additive to enhance phosphopeptide signals in MALDI-TOF-MS (Kjellstrom & Jensen, 2004). In both positive andnegative ionization modes, when 200 fmol of a-casein wasanalyzed, the use of phosphoric acid in the matrix solutionyielded increased signals from all four observed phosphopep-tides. In analysis of 100 fmol ofb-casein digest using positive-ionmode, the ionization of the tetraphosphorylated peptide was also

enhanced in the presence of phosphoric acid (Kjellstrom &Jensen, 2004; Stensballe & Jensen, 2004). This tetraphosphory-lated peptide was not detected when TFA, rather than phosphoricacid, was used in the matrix solution. The authors suggest that thephosphoric acid-enhanced detection of phosphorylated peptidesis due to a ‘‘salting out’’ effect, as PO4

3� is known to be the bestanion for precipitating proteins. Both 2,5-DHB and a-CHCAare widely used in phosphopeptide analysis, but in our hands 2,5-DHB/phosphoric acid has typically yielded stronger phospho-peptide signals than a-CHCA/diammonium hydrogen citrate.(Asara and Allison introduced ammonium salts includingdiammonium hydrogen citrate and ammonium acetate to theMALDI matrix for enhanced detection of phosphorylatedpeptides using positive-ion mode MALDI-TOF-MS; Asara &Allison, 1999.) Under the conditions we examined, ‘‘sweetspots’’ were easier to find in a 2,5-DHB/phosphoric acid matrixthan in a THAP/DAHC matrix.

V. PERSPECTIVE

Immobilized metal affinity chromatography (IMAC), MOAC,and covalent methods are all capable of selectively enrichingphosphopeptides, and thus far one technique has not emerged assuperior to the others. Covalent techniques are sometimes time-consuming with many steps, but several recent studies presentedsimplified procedures (Tao et al., 2005; Bodenmiller et al.,2007b). Still, esterification of carboxylic acid groups is usuallyrequired for these techniques, and if the esterification is less thanquantitative, this will complicate mass spectra (Yu et al., 2007;Simon et al., 2008). As mentioned previously, there will also bepeptide loss during lyophilization after esterification (Stewart,Thomson, & Figeys, 2001).

Metal oxide affinity chromatography (MOAC) based onadsorption to TiO2 is especially attractive, but as with alltechniques, loading, rinsing, and elution solutions must becarefully selected to minimize nonspecific adsorption and tomaximize the detection of both monophosphorylated andmultiphosphorylated species. It appears that with appropriatereagents, it might not be necessary to esterify proteins prior totheir enrichment on TiO2. However, recent work by Simon et al.suggests that methyl esterification does enhance the specificity ofthe technique when 2,5-DHB is used to exclude some nonspecificadsorption (Simon et al., 2008). Recently developed ‘‘excluders’’such as phthalic acid (Thingholm et al., 2006), glycolic acid(Jensen & Larsen, 2007), and ammonium glutamate (Yu et al.,2007) seem to allow purification without the drawbacks of methylesterification. Glycolic acid, as well as some of the other newexcluders, may be more compatible with LC-MS than 2,5-DHB.

Immobilized metal affinity chromatography (IMAC) mightnot provide the selectivity available with TiO2 enrichment, butwith appropriate reagents, IMAC can be selective and sensitivefor monophosphorylated and tetraphosphorylated peptides.However, some buffers and reagents such as EDTA are notcompatible with IMAC, so HPLC purification may be neededprior to this technique (Jensen & Larsen, 2007). MOAC appearsto be more tolerant to EDTA and other reagents such assurfactants.



One challenge to assess the utility of different techniques isthe limited pool of proteins that are examined in model systems.Many studies have focused on a- and b-casein because of the lowcost of these proteins. Development of a more complete series ofphosphopeptides with a wider range of residue compositionswould enhance our understanding of the efficiency of phospho-peptide enrichment using different systems. Additionally, thehigh recoveries achieved in simple protein digests are most likelynot relevant to more complex samples. However, evaluation ofcell lysates is challenging because the phosphopeptide compo-sition is unknown.

A combination of enrichment techniques is needed toachieve the highest phosphopeptide coverage. Bodenmiller andco-workers recently compared IMAC, MOAC (TiO2 usingphthalic acid as an excluder), and covalent phosphoramidateenrichment using the same digest of the cytosolic fraction ofDrosophila melanogaster Kc167 cells (Bodenmiller et al.,2007b). After enrichment, phosphopeptides were identified withLC-ESI tandem mass spectrometry. Although the study could notsay what fraction of the total phosphopeptides were identified, itdid allow an important comparison of the phosphopeptidesidentified by the different techniques. IMAC, MOAC, andcovalent enrichment identified 366, 535, and 555 phosphopep-tides, respectively, but the overlap of phosphopeptides identifiedusing the different techniques was only approximately 34%.Thus, although all of the methods showed relatively highspecificity for phosphopeptides, a combination of techniques isneeded for the highest phosphopeptide identification. Overall,nearly 900 phosphopeptides were identified using the combina-tion of the three techniques. Nevertheless, this may still be only asmall fraction of the total phosphopeptides in the cell. Spiking ofcell lysates with a series of 10–20 synthetic phosphopeptidesmight provide a useful system for investigating the level at whichphosphopeptides can be detected as well as the comprehensive-ness of the methods.

The choice of enrichment platform (column, pipette tip,magnetic bead, or modified MALDI plate) will likely depend onthe particular application. When trying to isolate and identify asmany phosphoproteins as possible in a cell lysate, chromato-graphic column-based methods are required. Multiple elutionsfrom IMAC or MOAC columns or even gradient elutions can helpto simplify fractions of proteins and reveal more peptides (Simonet al., 2008; Thingholm et al., 2008). However, as mentionedabove, even with a combination of methods, it is not clear whatfraction of phosphopeptides can be identified in a complexsample.

For heterogeneous samples, extraction with magnetic beadsis attractive to avoid the need for filtration prior to enrichment. Inthe case of small (mL), relatively pure samples (e.g., fromimmunoprecipitated proteins) pipette tips and modified MALDIplates are both attractive. However, the on-plate enrichmentoffers advantages in terms of simplicity and minimal sample loss.With detection limits at the femtomole level, modified plates arebecoming more viable for such applications. However, thesetechniques are not likely to prove suitable for complex samples.Additionally, it is important to carefully select the capacity ofthe extractant (magnetic bead, plate, or column) because whenthe binding capacity of the material is exceeded, there will beselective extraction of specific phosphopeptides, frequently

multiply phosphorylated species. In contrast with too high abinding capacity, nonspecific interactions might occur. Phos-phopeptide enrichment using magnetic beads and modifiedMALDI plates is still relatively new and has not yet gainedwidespread use.

Overall, the potential for phosphopeptide enrichment andrapid analysis has improved dramatically over the past 5 years.A combination of techniques can reveal large numbers ofphosphopeptides in complex samples, but comprehensivephosphoproteomics is still not possible. For the highest proteincoverage, future phosphoproteomic techniques will likelyemploy multiple enrichment techniques along with two-dimen-sional separations, but such studies are time consuming. Thecombination of immunoprecipitation and enrichment shouldprove useful for reducing the complexity of samples, but this willnot be comprehensive. At present the challenge of completephosphoproteomics is daunting at best. Comparisons of the levelsof specific phosphopeptides/phosphoproteins in different cells orin response to specific treatments are much more feasible. In suchcases, simple techniques such as on-plate purification can help torapidly identify phosphorylation sites in proteins that are isolatedfrom complex mixtures. We expect that many future studies willfocus on the use of these techniques to understand phosphory-lation states in important biochemical pathways.

REFERENCES

Adamczyk M, Gebler JC, Wu J. 2001. Selective analysis of phosphopeptides

within a protein mixture by chemical modification, reversible

biotinylation and mass spectrometry. Rapid Commun Mass Spectrom

15:1481–1488.

Adams JA. 2001. Kinetic and catalytic mechanisms of protein kinases. Chem

Rev 101:2271–2290.

Aebersold R, Goodlett DR. 2001. Mass spectrometry in proteomics. Chem

Rev 101:269–295.

Akashi T, Yamori T. 2007. A novel method for analyzing phosphoproteins

using SELDI-TOF MS in combination with a series of recombinant

proteins. Proteomics 7:2350–2354.

Akashi T, Nishimura Y, Wakatabe R, Shiwa M, Yamori T. 2007. Proteomics-

based identification of biomarkers for predicting sensitivity to a

PI3-kinase inhibitor in cancer. Biochem Biophys Res Commun

352:514–521.

Andersson L, Porath J. 1986. Isolation of phosphoproteins by immobilized

metal (Fe3þ) affinity-chromatography. Anal Biochem 154:250–254.

Asara JM, Allison J. 1999. Enhanced detection of phosphopeptides in matrix-

assisted laser desorption/ionization mass spectrometry using ammo-

nium salts. J Am Soc Mass Spectrom 10:35–44.

Ataka K, Heberle J. 2006. Use of surface enhanced infrared absorption

spectroscopy (SEIRA) to probe the functionality of a protein

monolayer. Biopolymers 82:415–419.

Ataka K, Richter B, Heberle J. 2006. Orientational control of the

physiological reaction of cytochrome c oxidase tethered to a gold

electrode. J Phys Chem B 110:9339–9347.

Ballif BA, Villen J, Beausoleil SA, Schwartz D, Gygi SP. 2004.

Phosphoproteomic analysis of the developing mouse brain. Mol Cell

Proteomics 3:1093–1101.

Beausoleil SA, Jedrychowski M, Schwartz D, Elias JE, Villen J, Li JX, Cohn

MA, Cantley LC, Gygi SP. 2004. Large-scale characterization of HeLa

cell nuclear phosphoproteins. Proc Natl Acad Sci USA 101:12130–

12135.



Belisle CM, Chen IY, Mirshad JK, Roth U, Steinert K, John A, Walker I. 54th

ASMS Conference on Mass Spectrometry, Seattle, WA, May 28–June

1, 2006.

Benitez IO, Bujoli B, Camus LJ, Lee CM, Odobel F, Talham DR. 2002.

Monolayers as models for supported catalysts: Zirconium phosphonate

films containing manganese(III) porphyrins. J Am Chem Soc

124:4363–4370.

Beranova-Giorgianni S, Zhao Y, Desiderio DM, Giorgianni F. 2006.

Phosphoproteomic analysis of the human pituitary. Pituitary 9:109–

120.

Blacken GR, Volny M, Vaisar T, Sadilek M, Turecek F. 2007. In situ

enrichment of phosphopeptides on MALDI plates functionalized by

reactive landing of zirconium(IV)-n-propoxide ions. Anal Chem

79:5449–5456.

Blackwell JA, Carr PW. 1991a. Fluoride-modified zirconium-oxide as a

biocompatible stationary phase for high-performance liquid-chroma-

tography. J Chromatogr 549:59–75.

Blackwell JA, Carr PW. 1991b. Study of the fluoride adsorption character-

istics of porous microparticulate zirconium-oxide. J Chromatogr 549:

43–57.

Bodega G, Suarez I, Almonacid L, Ciordia S, Beloso A, Lopez-Fernandez

LA, Zaballos A, Fernandez B. 2007. Effect of ammonia on ciliary

neurotrophic factor mRNA and protein expression and its upstream

signalling pathway in cultured rat astroglial cells: Possible implication

of c-fos, Sp1 and p38MAPK. Neuropathol Appl Neurobiol 33:420–

430.

Bodenmiller B, Mueller LN, Mueller M, Domon B, Aebersold R. 2007a.

Reproducible isolation of distinct, overlapping segments of the

phosphoproteome. Nat Methods 4:231–237.

Bodenmiller B, Mueller LN, Pedrioli PGA, Pflieger D, Junger MA, Eng JK,

Aebersold R, Tao WA. 2007b. An integrated chemical, mass

spectrometric and computational strategy for (quantitative) phospho-

proteomics: Application to drosophila melanogaster kc167 cells. Mol

Biosyst 3:275–286.

Bowley E, Mulvihill E, Howard JC, Pak BJ, Gan BS, Gorman DBO. 2005. A

novel mass spectrometry-based assay for GSK-3b activity. BMC

Biochem 6:29.

Brockman AH, Orlando R. 1995. Probe immobilized affinity-chromatog-

raphy mass-spectrometry. Anal Chem 67:4581–4585.

Brockman AH, Orlando R. 1996. New immobilization chemistry for probe

affinity mass spectrometry. Rapid Commun Mass Spectrom 10:1688–

1692.

Byrd H, Pike JK, Talham DR. 1993. Inorganic monolayers formed at an

organic template—A langmuir-blodgett route to monolayer and

multilayer films of zirconium octadecylphosphonate. Chem Mater

5:709–715.

Byrd H, Whipps S, Pike JK, Ma JF, Nagler SE, Talham DR. 1994. Role of the

template layer in organizing self-assembled films—Zirconium phos-

phonate monolayers and multilayers at a langmuir-blodgett template.

J Am Chem Soc 116:295–301.

Cantin GT, Shock TR, Park SK, Madhani HD, Yates JR. 2007. Optimizing

TiO2-based phosphopeptide enrichment for automated multidimen-

sional liquid chromatography coupled to tandem mass spectrometry.

Anal Chem 79:4666–4673.

Cardone MH, Roy N, Stennicke HR, Salvesen GS, Franke TF, Stanbridge E,

Frisch S, Reed JC. 1998. Regulation of cell death protease caspase-9 by

phosphorylation. Science 282:1318–1321.

Chen CT, Chen YC. 2005. Fe3O4/TiO2 core/shell nanoparticles as affinity

probes for the analysis of phosphopeptides using TiO2 surface-assisted

laser desorption/ionization mass spectrometry. Anal Chem 77:5912–

5919.

Chen CT, Chen YC. 2008. Trapping performance of Fe3O4 @ Al2O3 and

Fe3O4 @ TiO2 magnetic nanoparticles in the selective enrichment of

phosphopeptides from human serum. J Biomed Nanotechnol 4:73–79.

Chen CT, Chen WY, Tsai PJ, Chien KY, Yu JS, Chen YC. 2007. Rapid

enrichment of phosphopeptides and phosphoproteins from complex

samples using magnetic particles coated with alumina as the

concentrating probes for MALDI MS analysis. J Proteome Res 6:

316–325.

Clearfield A, Poojary DM, Zhang BL, Zhao BY, Derecskei-Kovacs A. 2000.

Azacrown ether pillared layered zirconium phosphonates and the

crystal structure of N,N0-bis(phosphonomethyl)-1,10-diaza-18-crown-

6. Chem Mater 12:2745–2752.

Clipston NL, Jai-nhuknan J, Cassady CJ. 2003. A comparison of negative and

positive ion time-of-flight post-source decay mass spectrometry for

peptides containing basic residues. Int J Mass Spectrom 222:363–381.

Cohen P. 2001. The role of protein phosphorylation in human health and

disease—Delivered on June 30th 2001 at the FEBS meeting in Lisbon.

Eur J Biochem 268:5001–5010.

Connor PA, McQuillan AJ. 1999. Phosphate adsorption onto TiO2 from

aqueous solutions: An in situ internal reflection infrared spectroscopic

study. Langmuir 15:2916–2921.

Craig AG, Hoeger CA, Miller CL, Goedken T, Rivier JE, Fischer WH. 1994.

Monitoring protein-kinase and phosphatase reactions with matrix-

assisted laser desorption/ionization mass spectrometry and capillary

zone electrophoresis—Comparison of the detection efficiency of

peptide-phosphopeptide mixtures. Biol Mass Spectrom 23:519–528.

Cuccurullo M, Schlosser G, Cacace G, Malorni L, Pocsfalvi G. 2007.

Identification of phosphoproteins and determination of phosphorylation

sites by zirconium dioxide enrichment and SELDI-MS/MS. J Mass

Spectrom 42:1069–1078.

D’Ambrosio C, Salzano AM, Arena S, Renzone G, Scaloni A. 2007.

Analytical methodologies for the detection and structural character-

ization of phosphorylated proteins. J Chromatogr B Anal Technol

Biomed Life Sci 849:163–180.

Davies HA. 2000. The ProteinChip System from Ciphergen: A new technique

for rapid, micro-scale protein biology. J Mol Med 78:B29–29.

Diella F, Cameron S, Gemund C, Linding R, Via A, Kuster B, Sicheritz-

Ponten T, Blom N, Gibson TJ. 2004. Phospho.Elm: A database of

experimentally verified phosphorylation sites in eukaryotic proteins.

BMC Bioinformatics 5:79.

Dobson KD, McQuillan AJ. 2000. In situ infrared spectroscopic analysis of

the adsorption of aromatic carboxylic acids to TiO2, ZrO2, Al2O3, and

Ta2O5 from aqueous solutions. Spectrochim Acta A Mol Biomol

Spectrosc 56:557–565.

Dogruel D, Williams P, Nelson RW. 1995. Rapid tryptic mapping using

enzymatically active mass-spectrometer probe tips. Anal Chem 67:

4343–4348.

Dong J, Zhou H, Wu R, Ye M, Zou H. 2007. Specific capture of

phosphopeptides by Zr4þ-modified monolithic capillary column.

J Sep Sci 30:2917–2923.

Dunn JD, Watson JT, Bruening ML. 2006. Detection of phosphopeptides

using Fe(III)-nitrilotriacetate complexes immobilized on a MALDI

plate. Anal Chem 78:1574–1580.

Dunn JD, Igrisan EA, Palumbo AM, Reid GE. 2008. Phosphopeptide

enrichment using MALDI plates modified with high-capacity polymer

brushes. Anal Chem 80:5727–5735.

Ekstrom S, Wallman L, Malm J, Becker C, Lilja H, Laurell T, Marko-Varga G.

2004. Integrated selective enrichment target—A microtechnology

platform for matrix-assisted laser desorption/ionization-mass spec-

trometry applied on protein biomarkers in prostate diseases. Electro-

phoresis 25:3769–3777.

Ekstrom S, Wallman L, Hok D, Marko-Varga G, Laurell T. 2006.

Miniaturized solid-phase extraction and sample preparation for MALDI

MS using a microfabricated integrated selective enrichment target. J

Proteome Res 5:1071–1081.

Ekstrom S, Wallman L, Helldin G, Nilsson J, Marko-Varga G, Laurell T. 2007.

Polymeric integrated selective enrichment target (ISET) for solid-



phase-based sample preparation in MALDI-TOF MS. J Mass Spectrom

42:1445–1452.

Espina V, Dettloff KA, Cowherd S, Petricoin EF, Lotta LA. 2004. Use of

proteomic analysis to monitor responses to biological therapies. Expert

Opin Biol Ther 4:83–93.

Feng S, Ye ML, Zhou HJ, Jiang XG, Jiang XN, Zou HF, Gong BL. 2007.

Immobilized zirconium ion affinity chromatography for specific

enrichment of phosphopeptides in phosphoproteome analysis. Mol

Cell Proteomics 6:1656–1665.

Ficarro SB, McCleland ML, Stukenberg PT, Burke DJ, Ross MM,

Shabanowitz J, Hunt DF, White FM. 2002. Phosphoproteome analysis

by mass spectrometry and its application to saccharomyces cerevisiae.

Nat Biotechnol 20:301–305.

Ficarro SB, Parikh JR, Blank NC, Marto JA. 2008. Niobium(V) oxide (Nb2O5):

Application to phosphoproteomics. Anal Chem 80:4606–4613.

Fischer EH, Krebs EG. 1955. Conversion of phosphorylase-b to phosphor-

ylase-a in muscle extracts. J Biol Chem 216:121–132.

Frens G. 1973. Controlled nucleation for regulation of particle-size in

monodisperse gold suspensions. Nat Phys Sci 241:20–22.

Frey BL, Hanken DG, Corn RM. 1993. Vibrational spectroscopic studies of

the attachment chemistry for zirconium phosphonate multilayers at gold

and germanium surfaces. Langmuir 9:1815–1820.

Gafken PR, Lampe PD. 2006. Methodologies for characterizing phospho-

proteins by mass spectrometry. Cell Commun Adhes 13:249–262.

Gamsjaeger R, Wimmer B, Kahr H, Tinazli A, Picuric S, Lata S, Tampe R,

Maulet Y, Gruber HJ, Hinterdorfer P, Romanin C. 2004. Oriented

binding of the His6-tagged carboxyl-tail of the L-type Ca2þ channel a1-

subunit to a new NTA-functionalized self-assembled monolayer.

Langmuir 20:5885–5890.

Ge Y, Gibbs BF, Masse R. 2005. Complete chemical and enzymatic treatment

of phosphorylated and glycosylated proteins on proteinchip arrays.

Anal Chem 77:3644–3650.

Giorgianni F, Zhao YX, Desiderio DM, Beranova-Giorgianni S. 2007.

Toward a global characterization of the phosphoproteome in prostate

cancer cells: Identification of phosphoproteins in the lncap cell line.

Electrophoresis 28:2027–2034.

Goshe MB, Veenstra TD, Panisko EA, Conrads TP, Angell NH, Smith RD.

2002. Phosphoprotein isotope-coded affinity tags: Application to the

enrichment and identification of low-abundance phosphoproteins. Anal

Chem 74:607–616.

Graves JD, Krebs EG. 1999. Protein phosphorylation and signal transduction.

Pharmacol Ther 82:111–121.

Grønborg M, Kristiansen TZ, Stensballe A, Andersen JS, Ohara O, Mann M,

Jensen ON, Pandey A. 2002. A mass spectrometry-based proteomic

approach for identification of serine/threonine-phosphorylated proteins

by enrichment with phospho-specific antibodies—Identification of a

novel protein, Frigg, as a protein kinase a substrate. Mol Cell


Gruhler A, Olsen JV, Mohammed S, Mortensen P, Faergeman NJ, Mann M,

Jensen ON. 2005. Quantitative phosphoproteomics applied to the yeast

pheromone signaling pathway. Mol Cell Proteomics 4:310–327.

Han GH, Ye ML, Zou HF. 2008. Development of phosphopeptide enrichment

techniques for phosphoproteome analysis. Analyst 133:1128–

1138.

Hart SR, Waterfield MD, Burlingame AL, Cramer R. 2002. Factors governing

the solubilization of phosphopeptides retained on ferric NTA IMAC

beads and their analysis by MALDI TOFMS. J Am Soc Mass Spectrom

13:1042–1051.

Head E, Moffat K, Das P, Sarsoza E, Poon WW, Landsberg G, Cotman CW,

Murphy MP. 2005. b-Amyloid deposition and tau phosphorylation in

clinically characterized aged cats. Neurobiol Aging 26:749–763.

Hochuli E, Dobeli H, Schacher A. 1987. New metal chelate adsorbent

selective for proteins and peptides containing neighboring histidine

residues. J Chromatogr 411:177–184.

Holmes LD, Schiller MR. 1997. Immobilized iron(III) metal affinity

chromatography for the separation of phosphorylated macromolecules:

Ligands and applications. J Liq Chromatogr Relat Technol 20:123–

142.

Hong HG, Sackett DD, Mallouk TE. 1991. Adsorption of well-ordered

zirconium phosphonate multilayer films on high surface-area silica.

Chem Mater 3:521–527.

Hubbard MJ, Cohen P. 1993. On target with a new mechanism for the regulation

of protein phosphorylation. Trends Biochem Sci 18:172–177.

Hunter T. 2000. Signaling—2000 and beyond. Cell 100:113–127.

Hunter T, Sefton BM. 1980. Transforming gene-product of rous-sarcoma

virus phosphorylates tyrosine. Proc Natl Acad Sci USA 77:1311–1315.

Hutchens TW, Yip TT. 1993. New desorption strategies for the mass-

spectrometric analysis of macromolecules. Rapid Commun Mass

Spectrom 7:576–580.

Ibanez AJ, Muck A, Svatos A. 2007. Metal-chelating plastic MALDI

(pMALDI) chips for the enhancement of phosphorylated-peptide/

protein signals. J Proteome Res 6:3842–3848.

Ikeguchi Y, Nakamura H. 1997. Determination of organic phosphates by

column-switching high performance anion-exchange chromatography

using on-line preconcentration on titania. Anal Sci 13:479–483.

Ikeguchi Y, Nakamura H. 2000. Selective enrichment of phospholipids by

titania. Anal Sci 16:541–543.

Jensen SS, Larsen MR. 2007. Evaluation of the impact of some experimental

procedures on different phosphopeptide enrichment techniques. Rapid

Commun Mass Spectrom 21:3635–3645.

Johnson SA, Hunter T. 2005. Kinomics: Methods for deciphering the kinome.

Nat Methods 2:17–25.

Johnson LN, Lewis RJ. 2001. Structural basis for control by phosphorylation.

Chem Rev 101:2209–2242.

Johnson DL, Martin LL. 2005. Controlling protein orientation at interfaces

using histidine tags: An alternative to Ni/NTA. J Am Chem Soc

127:2018–2019.

Kada G, Riener CK, Hinterdorfer P, Kienberger F, Stoh CM, Gruber HJ. 2002.

Dithio-phospholipids for biospecific immobilization of proteins on gold

surfaces. Single Mol 3:119–125.

Kagedal L. 1998. Immobilized metal ion affinity chromatography. In:

Janson J-C, Ryden L, editors. Protein purification: Principles, high-

resolution methods, and applications. New York: John Wiley and Sons,

pp 311–342.

Kjellstrom S, Jensen ON. 2004. Phosphoric acid as a matrix additive for

MALDI MS analysis of phosphopeptides and phosphoproteins. Anal

Chem 76:5109–5117.

Klenkar G, Valiokas R, Lundstrom I, Tinazli A, Tampe R, Piehler J, Liedberg

B. 2006. Piezo dispensed microarray of multivalent chelating thiols for

dissecting complex protein-protein interactions. Anal Chem 78:3643–

3650.

Kohli P, Blanchard GJ. 2000. Probing interfaces and surface reactions of

zirconium phosphate/phosphonate multilayers using 31P NMR spec-

trometry. Langmuir 16:695–701.

Koizumi Y, Taya M. 2002. Kinetic evaluation of biocidal activity of titanium

dioxide against phage MS2 considering interaction between the phage

and photocatalyst particles. Biochem Eng J 12:107–116.

Kokubu M, Ishihama Y, Sato T, Nagasu T, Oda Y. 2005. Specificity of

immobilized metal affinity-based IMAC/C18 tip enrichment of

phosphopeptides for protein phosphorylation analysis. Anal Chem

77:5144–5154.

Kosmulski M. 2002. The significance of the difference in the point of zero

charge between rutile and anatase. Adv Colloid Interface Sci 99:255–

264.

Krause E, Wenschuh H, Jungblut PR. 1999. The dominance of arginine-

containing peptides in MALDI-derived tryptic mass fingerprints of

proteins. Anal Chem 71:4160–4165.



Krebs EG. 1983. Historical perspectives on protein-phosphorylation and a

classification-system for protein-kinases. Philos Trans Roy Soc B 302:

3–11.

Krebs EG, Beavo JA. 1979. Phosphorylation-dephosphorylation of enzymes.

Annu Rev Biochem 48:923–959.

Kristjansdottir K, Wolfgeher D, Lucius N, Angulo DS, Kron SJ. 2008.

Phosphoprotein profiling by PA-GeLC-MS/MS. J Proteome Res

7:2812–2824.

Kumar CV, Chaudhari A. 2000. Proteins immobilized at the galleries of

layered a-zirconium phosphate: Structure and activity studies. J Am

Chem Soc 122:830–837.

Kweon HK, Hakansson K. 2006. Selective zirconium dioxide-based enrich-

ment of phosphorylated peptides for mass spectrometric analysis. Anal

Chem 78:1743–1749.

Laine J, Kunstle G, Obata T, Sha M, Noguchi M. 2000. The protooncogene

TCL1 is an Akt kinase coactivator. Mol Cell 6:395–407.

Lansdell TA, Tepe JJ. 2004. Isolation of phosphopeptides using solid phase

enrichment. Tetrahedron Lett 45:91–93.

Larsen MR, Thingholm TE, Jensen ON, Roepstorff P, Jorgensen TJD. 2005.

Highly selective enrichment of phosphorylated peptides from peptide

mixtures using titanium dioxide microcolumns. Mol Cell Proteomics

4:873–886.

Le Bihan MC, Hou YW, Harris N, Tarelli E, Coulton GR. 2006. Proteomic

analysis of fast and slow muscles from normal and kyphoscoliotic mice

using protein arrays, 2-DE and MS. Proteomics 6:4646–4661.

Lee H, Kepley LJ, Hong HG, Akhter S, Mallouk TE. 1988. Adsorption of

ordered zirconium phosphonate multilayer films on silicon and gold

surfaces. J Phys Chem 92:2597–2601.

Lee JK, Kim YG, Chi YS, Yun WS, Choi IS. 2004. Grafting nitrilotriacetic

groups onto carboxylic acid-terminated self-assembled monolayers on

gold surfaces for immobilization of histidine-tagged proteins. J Phys

Chem B 108:7665–7673.

Li SH, Dass C. 1999. Iron(lll)-immobilized metal ion affinity chromatog-

raphy and mass spectrometry for the purification and characterization of

synthetic phosphopeptides. Anal Biochem 270:9–14.

Li Y, Leng TH, Lin HQ, Deng CH, Xu XQ, Yao N, Yang PY, Zhang XM.

2007a. Preparation of Fe3O4@ZrO2 core-shell microspheres as affinity

probes for selective enrichment and direct determination of phospho-

peptides using matrix-assisted laser desorption ionization mass

spectrometry. J Proteome Res 6:4498–4510.

Li Y, Liu Y, Tang J, Lin H, Yao N, Shen X, Deng C, Yang P, Zhang X. 2007b.

Fe3O4@Al2O3 magnetic core-shell microspheres for rapid and highly

specific capture of phosphopeptides with mass spectrometry analysis. J

Chromatogr A 1172:57–71.

Li Y, Lin HQ, Deng CH, Yang PY, Zhang XM. 2008. Highly selective and

rapid enrichment of phosphorylated peptides using gallium oxide-

coated magnetic microspheres for MALDI-TOF-MS and nano-LC-

ESI-MS/MS/MS analysis. Proteomics 8:238–249.

Liang SS, Makamba H, Huang SY, Chen SH. 2006. Nano-titanium dioxide

composites for the enrichment of phosphopeptides. J Chromatogr A

1116:38–45.

Liang XQ, Fonnum G, Hajivandi M, Stene T, Kjus NH, Ragnhildstveit E,

Arnshey JW, Predki P, Pope RM. 2007. Quantitative comparison of

IMAC and TiO2 surfaces used in the study of regulated, dynamic protein

phosphorylation. J Am Soc Mass Spectrom 18:1932–1944.

Liao PC, Leykam J, Andrews PC, Gage DA, Allison J. 1994. An approach to

locate phosphorylation sites in a phosphoprotein—Mass mapping by

combining specific enzymatic degradation with matrix-assisted laser-

desorption ionization mass-spectrometry. Anal Biochem 219:9–20.

Lim YP. 2005. Mining the tumor phosphoproteome for cancer markers. Clin

Cancer Res 11:3163–3169.

Lim KB, Kassel DB. 2006. Phosphopeptides enrichment using on-line two-

dimensional strong cation exchange followed by reversed-phase liquid

chromatography/mass spectrometry. Anal Biochem 354:213–219.

Lin HY, Chen CT, Chen YC. 2006. Detection of phosphopeptides by localized

surface plasma resonance of titania-coated gold nanoparticles immo-

bilized on glass substrates. Anal Chem 78:6873–6878.

Liu Y, Porta A, Peng XR, Gengaro K, Cunningham EB, Li H, Dominguez LA,

Bellido T, Christakos S. 2004. Prevention of glucocorticoid-induced

apoptosis in osteocytes and osteoblasts by calbindin-D28k. J Bone Miner

Res 19:479–490.

Lo CY, Chen WY, Chen CT, Chen YC. 2007. Rapid enrichment of

phosphopeptides from tryptic digests of proteins using iron oxide

nanocomposites of magnetic particles coated with zirconia as the

concentrating probes. J Proteome Res 6:887–893.

Luk YY, Tingey ML, Hall DJ, Israel BA, Murphy CJ, Bertics PJ, Abbott NL.

2003. Using liquid crystals to amplify protein-receptor interactions:

Design of surfaces with nanometer-scale topography that present

histidine-tagged protein receptors. Langmuir 19:1671–1680.

Makower A, Halamek J, Skladal P, Kernchen F, Scheller FW. 2003. New

principle of direct real-time monitoring of the interaction of

cholinesterase and its inhibitors by piezolectric biosensor. Biosens

Bioelectron 18:1329–1337.

Maly J, Di Meo C, De Francesco M, Masci A, Masojidek J, Sugiura M, Volpe

A, Pilloton R. 2004a. Reversible immobilization of engineered

molecules by Ni-NTA chelators. Bioelectrochemistry 63:271–275.

Maly J, Masci A, Masojidek J, Sugiura M, Pilloton R. 2004b. Monolayers of

natural and recombinant photosystem II on gold electrodes—Potentials

for use as biosensors for detection of herbicides. Anal Lett 37:1645–

1656.

Mann M, Ong SE, Grønborg M, Steen H, Jensen ON, Pandey A. 2002.

Analysis of protein phosphorylation using mass spectrometry: Deci-

phering the phosphoproteome. Trends Biotechnol 20:261–268.

Matsuoka S, Ballif BA, Smogorzewska A, McDonald ER, Hurov KE, Luo J,

Bakalarski CE, Zhao ZM, Solimini N, Lerenthal Y, Shiloh Y, Gygi SP,

Elledge SJ. 2007. ATM and ATR substrate analysis reveals extensive

protein networks responsive to DNA damage. Science 316:1160–1166.

Mazur M, Krysinski P, Blanchard GJ. 2005. Use of zirconium-phosphate-

carbonate chemistry to immobilize polycyclic aromatic hydrocarbons

on boron-doped diamond. Langmuir 21:8802–8808.

Morandell S, Stasyk T, Grosstessner-Hain K, Roitinger E, Mechtler K, Bonn

GK, Huber LA. 2006. Phosphoproteomics strategies for the functional

analysis of signal transduction. Proteomics 6:4047–4056.

Mortier E, Cornelissen F, van Hove C, Dillen L, Richardson A. 2001. The

focal adhesion targeting sequence is the major inhibitory moiety of Fak-

related non-kinase. Cell Signal 13:901–909.

Moser K, White FM. 2006. Phosphoproteomic analysis of rat liver by high

capacity IMAC and LC-MS/MS. J Proteome Res 5:98–104.

Nawrocki J, Dunlap C, McCormick A, Carr PW. 2004. Part I. Chromatog-

raphy using ultra-stable metal oxide-based stationary phases for HPLC.

J Chromatogr A 1028:1–30.

Ndassa YM, Orsi C, Marto JA, Chen S, Ross MM. 2006. Improved

immobilized metal affinity chromatography for large-scale phospho-

proteomics applications. J Proteome Res 5:2789–2799.

Nelson RW, Dogruel D, Krone JR, Williams P. 1995. Peptide characterization

using bioreactive mass-spectrometer probe tips. Rapid Commun Mass

Spectrom 9:1380–1385.

Neville DCA, Rozanas CR, Price EM, Gruis DB, Verkman AS, Townsend RR.

1997. Evidence for phosphorylation of serine 753 in cftr using a novel

metal-ion affinity resin and matrix-assisted laser desorption mass

spectrometry. Protein Sci 6:2436–2445.

Nixon CN, Le Claire K, Odobel F, Bujoli B, Talham DR. 1999. Palladium

porphyrin containing zirconium phosphonate langmuir-blodgett films.

Chem Mater 11:965–976.

Nonglaton G, Benitez IO, Guisle I, Pipelier M, Leger J, Dubreuil D, Tellier C,

Talham DR, Bujoli B. 2004. New approach to oligonucleotide

microarrays using zirconium phosphonate-modified surfaces. J Am

Chem Soc 126:1497–1502.



Nousiainen M, Sillje HHW, Sauer G, Nigg EA, Korner R. 2006.

Phosphoproteome analysis of the human mitotic spindle. Proc Natl

Acad Sci USA 103:5391–5396.

Nuhse TS, Stensballe A, Jensen ON, Peck SC. 2003. Large-scale analysis of in

vivo phosphorylated membrane proteins by immobilized metal ion

affinity chromatography and mass spectrometry. Mol Cell Proteomics

2:1234–1243.

Oda Y, Nagasu T, Chait BT. 2001. Enrichment analysis of phosphorylated

proteins as a tool for probing the phosphoproteome. Nat Biotechnol

19:379–382.

Olsen JV, Blagoev B, Gnad F, Macek B, Kumar C, Mortensen P, Mann M.

2006. Global, in vivo, and site-specific phosphorylation dynamics in

signaling networks. Cell 127:635–648.

Papac DI, Hoyes J, Tomer KB. 1994. Direct analysis of affinity-bound

analytes by MALDI/TOF MS. Anal Chem 66:2609–2613.

Pawson T, Nash P. 2000. Protein-protein interactions define specificity in

signal transduction. Genes Dev 14:1027–1047.

Pinkse MWH, Uitto PM, Hilhorst MJ, Ooms B, Heck AJR. 2004. Selective

isolation at the femtomole level of phosphopeptides from proteolytic

digests using 2D-nanoLC-ESI-MS/MS and titanium oxide precolumns.

Anal Chem 76:3935–3943.

Porath J, Carlsson J, Olsson I, Belfrage G. 1975. Metal chelate affinity

chromatography, a new approach to protein fractionation. Nature

258:598–599.

Posewitz MC, Tempst P. 1999. Immobilized gallium(III) affinity chromatog-

raphy of phosphopeptides. Anal Chem 71:2883–2892.

Putvinski TM, Schilling ML, Katz HE, Chidsey CED, Mujsce AM, Emerson

AB. 1990. Self-assembly of organic multilayers with polar order using

zirconium-phosphate bonding between layers. Langmuir 6:1567–1571.

Qiao L, Roussel C, Wan JJ, Yang PY, Girault HH, Liu BH. 2007. Specific on-

plate enrichment of phosphorylated peptides for direct MALDI-TOF

MS analysis. J Proteome Res 6:4763–4769.

Reinders J, Sickmann A. 2005. State-of-the-art in phosphoproteomics.


Renger C, Kuschel P, Kristoffersson A, Clauss B, Oppermann W, Sigmund W.

2006. Adsorption studies on nano-zirconia in water and a water-1,2-

propanediol mixture. J Ceram Process Res 7:106–112.

Righetti PG, Castagna A, Antonucci F, Piubelli C, Cecconi D, Campostrini N,

Rustichelli C, Antonioli P, Zanusso G, Monaco S, Lomas L, Boschetti E.

2005. Proteome analysis in the clinical chemistry laboratory: Myth or

reality? Clin Chim Acta 357:123–139.

Rigler P, Ulrich WP, Hoffmann P, Mayer M, Vogel H. 2003. Reversible

immobilization of peptides: Surface modification and in situ detection

by attenuated total reflection FTIR spectroscopy. Chemphyschem

4:268–275.

Rigler P, Ulrich WP, Vogel H. 2004. Controlled immobilization of membrane

proteins to surfaces for Fourier transform infrared investigations.

Langmuir 20:7901–7903.

Rinalducci S, Larsen MR, Mohammed S, Zolla L. 2006. Novel protein

phosphorylation site identification in spinach stroma membranes by

titanium dioxide microcolumns and tandem mass spectrometry. J

Proteome Res 5:973–982.

Roig J, Tuazon PT, Zipfel PA, Pendergast AM, Traugh JA. 2000. Functional

interaction between c-Abl and the p21-activated protein kinase g-PAK.

Proc Natl Acad Sci USA 97:14346–14351.

Rush J, Moritz A, Lee KA, Guo A, Goss VL, Spek EJ, Zhang H, Zha XM,

Polakiewicz RD, Comb MJ. 2005. Immunoaffinity profiling of tyrosine

phosphorylation in cancer cells. Nat Biotechnol 23:94–101.

Sano A, Nakamura H. 2004a. Chemo-affinity of titania for the column-

switching HPLC analysis of phosphopeptides. Anal Sci 20:565–566.

Sano A, Nakamura H. 2004b. Titania as a chemo-affinity support for the

column-switching HPLC analysis of phosphopeptides: Application to

the characterization of phosphorylation sites in proteins by combination

with protease digestion and electrospray ionization mass spectrometry.

Anal Sci 20:861–864.

Scheibler L, Dumy P, Stamou D, Duschl C, Vogel H, Mutter M. 1998. Self-

assembling functionalized templates in biosensor technology. Polym

Bull 40:151–157.

Schmelzle K, White FM. 2006. Phosphoproteomic approaches to elucidate

cellular signaling networks. Curr Opin Biotechnol 17:406–414.

Schmidt SR, Schweikart F, Andersson ME. 2007. Current methods for

phosphoprotein isolation and enrichment. J Chromatogr B Anal

Technol Biomed Life Sci 849:154–162.

Seeley EH, Riggs LD, Regnier FE. 2005. Reduction of non-specific binding in

Ga(III) immobilized metal affinity chromatography for phosphopep-

tides by using endoproteinase glu-C as the digestive enzyme.

J Chromatogr B Anal Technol Biomed Life Sci 817:81–88.

Shen JW, Ahmed T, Vogt A, Wang JY, Severin J, Smith R, Dorwin S, Johnson

R, Harlan J, Holzman T. 2005. Preparation and characterization of

nitrilotriacetic-acid-terminated self-assembled monolayers on gold

surfaces for matrix-assisted laser desorption ionization-time of flight-

mass spectrometry analysis of proteins and peptides. Anal Biochem

345:258–269.

Sigal GB, Bamdad C, Barberis A, Strominger J, Whitesides GM. 1996. A self-

assembled monolayer for the binding and study of histidine tagged

proteins by surface plasmon resonance. Anal Chem 68:490–497.

Simon ES, Young M, Chan A, Bao ZQ, Andrews PC. 2008. Improved

enrichment strategies for phosphorylated peptides on titanium dioxide

using methyl esterification and pH gradient elution. Anal Biochem

377:234–242.

Simpson RJ. 2003. Proteins and proteomics: A laboratory manual. Cold

Spring Harbor: Cold Spring Harbor Laboratory Press.

Smith JC, Figeys D. 2008. Recent developments in mass spectrometry-based

quantitative phosphoproteomics. Biochem Cell Biol 86:137–

148.

Sopko R, Andrews BJ. 2008. Linking the kinome and phosphorylome—A

comprehensive review of approaches to find kinase targets. Mol Biosyst

4:920–933.

Steen H, Jebanathirajah JA, Rush J, Morrice N, Kirschner MW. 2006.

Phosphorylation analysis by mass spectrometry—Myths, facts, and the

consequences for qualitative and quantitative measurements. Mol Cell


Stensballe A, Jensen ON. 2004. Phosphoric acid enhances the performance of

Fe(III) affinity chromatography and matrix-assisted laser desorption/

ionization tandem mass spectrometry for recovery, detection and

sequencing of phosphopeptides. Rapid Commun Mass Spectrom

18:1721–1730.

Stewart II, Thomson T, Figeys D. 2001. O-18 labeling: A tool for proteomics.

Rapid Commun Mass Spectrom 15:2456–2465.

Stoica GE, Kuo A, Aigner A, Sunitha I, Souttou B, Malerczyk C, Caughey DJ,

Wen DZ, Karavanov A, Riegel AT, Wellstein A. 2001. Identification of

anaplastic lymphoma kinase as a receptor for the growth factor

pleiotrophin. J Biol Chem 276:16772–16779.

Stora T, Hovius R, Dienes Z, Pachoud M, Vogel H. 1997. Metal ion trace

detection by a chelator-modified gold electrode: A comparison of

surface to bulk affinity. Langmuir 13:5211–5214.

Sugiyama N, Masuda T, Shinoda K, Nakamura A, Tomita M, Ishihama Y.

2007. Phosphopeptide enrichment by aliphatic hydroxy acid-modified

metal oxide chromatography for nano-LC-MS/MS in proteomics

applications. Mol Cell Proteomics 6:1103–1109.

Tan F, Zhang YJ, Wang JL, Wei JY, Qin PB, Cai Y, Qian XH. 2007. Specific

capture of phosphopeptides on matrix-assisted laser desorption/

ionization time-of-flight mass spectrometry targets modified by

magnetic affinity nanoparticles. Rapid Commun Mass Spectrom 21:

2407–2414.

Tan F, Zhang Y, Mi W, Wang J, Wei J, Cai Y, Qian X. 2008. Enrichment of

phosphopeptides by Fe3þ-immobilized magnetic nanoparticles for



phosphoproteome analysis of the plasma membrane of mouse liver.

J Proteome Res 7:1078–1087.

Tang N, Tornatore P, Weinberger SR. 2004. Current developments in SELDI

affinity technology. Mass Spectrom Rev 23:34–44.

Tao WA, Wollscheid B, O’Brien R, Eng JK, Li XJ, Bodenmiller B, Watts JD,

Hood L, Aebersold R. 2005. Quantitative phosphoproteome analysis

using a dendrimer conjugation chemistry and tandem mass spectrom-

etry. Nat Methods 2:591–598.

Tassi E, Al-Attar A, Aigner A, Swift MR, McDonnell K, Karavanov A,

Wellstein A. 2001. Enhancement of fibroblast growth factor (FGF)

activity by an FGF-binding protein. J Biol Chem 276:40247–40253.

Thaler F, Valsasina B, Baldi R, Jin X, Stewart A, Isacchi A, Kalisz HM,

Rusconi L. 2003. A new approach to phosphoserine and phospho-

threonine analysis in peptides and proteins: Chemical modification,

enrichment via solid-phase reversible binding, and analysis by mass

spectrometry. Anal Bioanal Chem 376:366–373.

Thingholm TE, Jorgensen TJD, Jensen ON, Larsen MR. 2006. Highly

selective enrichment of phosphorylated peptides using titanium

dioxide. Nat Protoc 1:1929–1935.

Thingholm TE, Jensen ON, Robinson PJ, Larsen MR. 2008. SIMAC

(sequential elution from IMAC), a phosphoproteomics strategy for the

rapid separation of monophosphorylated from multiply phosphorylated

peptides. Mol Cell Proteomics 7:661–671.

Thulasiraman V, Wang Z, Katrekar A, Lomas L, Yip TT. 2004. Simultaneous

monitoring of multiple kinase activities by SELDI-TOF mass

spectrometry. In: Fung ET, editor. Protein arrays methods and protocol.

Totowa: Humana Press. pp 205–215.

Tinazli A, Tang JL, Valiokas R, Picuric S, Lata S, Piehler J, Liedberg B,

Tampe R. 2005. High-affinity chelator thiols for switchable and oriented

immobilization of histidine-tagged proteins: A generic platform for

protein chip technologies. Chem-A Eur J 11:5249–5259.

Trammell SA, Wang LY, Zullo JM, Shashidhar R, Lebedev N. 2004.

Orientated binding of photosynthetic reaction centers on gold using Ni-

NTA self-assembled monolayers. Biosens Bioelectron 19:1649–1655.

Trinidad JC, Specht CG, Thalhammer A, Schoepfer R, Burlingame AL. 2006.

Comprehensive identification of phosphorylation sites in postsynaptic

density preparations. Mol Cell Proteomics 5:914–922.

Valiokas R, Klenkar G, Tinazli A, Tampe R, Liedberg B, Piehler J. 2006.

Differential protein assembly on micropatterned surfaces with tailored

molecular and surface multivalency. Chembiochem 7:1325–1329.

Vassileva E, Proinova I, Hadjiivanov K. 1996. Solid-phase extraction of heavy

metal ions on a high surface area titanium dioxide (anatase). Analyst

121:607–612.

Vogel HJ. 1989. Phosphorus-31 nuclear magnetic resonance of phosphopro-

teins. Methods Enzymol 177:263–282.

Wang YB, Xia XH, Guo YL. 2005. Porous anodic alumina membrane as a

sample support for MALDI-TOF MS analysis of salt-containing

proteins. J Am Soc Mass Spectrom 16:1488–1492.

Wang YB, Chen W, Wu JS, Guo YL, Xia XH. 2007. Highly efficient and of

selective enrichment phosphopeptides using porous anodic alumina

membrane for MALDI-TOF MS analysis. J Am Soc Mass Spectrom

18:1387–1395.

Warthaka M, Karwowska-Desaulniers P, Pflum MKH. 2006. Phosphopeptide

modification and enrichment by oxidation-reduction condensation.

ACS Chem Biol 1:697–701.

Wegner GJ, Lee NJ, Marriott G, Corn RM. 2003. Fabrication of histidine-

tagged fusion protein arrays for surface plasmon resonance imaging

studies of protein-protein and protein-DNA interactions. Anal Chem

75:4740–4746.

Wei JY, Zhang YJ, Wang JL, Tan F, Liu JF, Cai Y, Qian XH. 2008. Highly

efficient enrichment of phosphopeptides by magnetic nanoparticles

coated with zirconium phosphonate for phosphoproteome analysis.

Rapid Commun Mass Spectrom 22:1069–1080.

Whitmarsh AJ, Davis RJ. 2000. Regulation of transcription factor function by

phosphorylation. Cell Mol Life Sci 57:1172–1183.

Wolschin F, Wienkoop S, Weckwerth W. 2005. Enrichment of phosphory-

lated proteins and peptides from complex mixtures using metal oxide/

hydroxide affinity chromatography (MOAC). Proteomics 5:4389–

4397.

Wu HFM, Jin M, Marsh CB. 2005. Toward functional proteomics of alveolar

macrophages. Am J Physiol-Lung Cell Mol Physiol 288:L585–

L595.

Wu J, Shakey Q, Liu W, Schuller A, Follettie MT. 2007. Global profiling of

phosphopeptides by titania affinity enrichment. J Proteome Res 6:

4684–4689.

Xie Y, Jiang Y, Ben-Amotz D. 2005. Detection of amino acid and peptide

phosphate protonation using Raman spectroscopy. Anal Biochem

343:223–230.

Xu YD, Watson JT, Bruening ML. 2003. Patterned monolayer/polymer films

for analysis of dilute or salt-contaminated protein samples by MALDI-

MS. Anal Chem 75:185–190.

Xu YD, Bruening ML, Watson JT. 2004. Use of polymer-modified MALDI-

MS probes to improve analyses of protein digests and DNA. Anal Chem

76:3106–3111.

Xu SY, Zhou HJ, Pan CS, Fu Y, Zhang Y, Li X, Ye ML, Zou HF. 2006.

Iminodiacetic acid derivatized porous silicon as a matrix support

for sample pretreatment and matrix-assisted laser desorption/

ionization time-of-flight mass spectrometry analysis. Rapid Commun

Mass Spectrom 20:1769–1775.

Yang XF, Wu HP, Kobayashi T, Solaro RJ, van Breemen RB. 2004. Enhanced

ionization of phosphorylated peptides during MALDI TOF mass

spectrometry. Anal Chem 76:1532–1536.

Yang F, Stenoien DL, Strittmatter EF, Wang JH, Ding LH, Lipton MS,

Monroe ME, Nicora CD, Gristenko MA, Tang KQ, Fang RH, Adkins

JN, Camp DG, Chen DJ, Smith RD. 2006. Phosphoproteome profiling

of human skin fibroblast cells in response to low- and high-dose

irradiation. J Proteome Res 5:1252–1260.

Yu L-R, Issaq HJ, Veenstra TD. 2007. Phosphoproteornics for the discovery of

kinases as cancer biomarkers and drug targets. Proteomics Clin Appl

1:1042–1057.

Yu L-R, Zhu Z, Chan KC, Issaq HJ, Dimitrov DS, Veenstra TD. 2007.

Improved titanium dioxide enrichment of phosphopeptides from HeLa

cells and high confident phosphopeptide identification by cross-

validation of MS/MS and MS/MS/MS spectra. J Proteome Res 6:

4150–4162.

Zhang GA, Neubert TA. 2006. Use of detergents to increase selectivity of

immunoprecipitation of tyrosine phosphorylated peptides prior to

identification by MALDI quadrupole-TOF MS. Proteomics 6:571–578.

Zhou HL, Watts JD, Aebersold R. 2001. A systematic approach to the analysis

of protein phosphorylation. Nat Biotechnol 19:375–378.

Zhou HJ, Xu SY, Ye ML, Feng S, Pan C, Jiang XG, Li X, Han GH, Fu Y, Zou

H. 2006. Zirconium phosphonate-modified porous silicon for highly

specific capture of phosphopeptides and MALDI-TOF MS analysis.

J Proteome Res 5:2431–2437.

Zhou HJ, Tian RJ, Ye ML, Xu SY, Feng S, Pan CS, Jiang XG, Li X, Zou HF.

2007. Highly specific enrichment of phosphopeptides by zirconium

dioxide nanoparticles for phosphoproteome analysis. Electrophoresis

28:2201–2215.



Jamie D. Dunn is currently a chemist of the U.S. Food and Drug Administration, Center for

Drug Evaluation and Research, Division of Pharmaceutical Analysis in St. Louis, MO. She

received her Ph.D. in Chemistry (2007) and M.S. in Chemistry and Criminal Justice from

Michigan State University. Her research interests include on-plate purification for MALDI-

MS and drug analysis using HPLC-MS.

Gavin E. Reid is an Assistant Professor in the Department of Chemistry and the Department

of Biochemistry and Molecular Biology at Michigan State University. He received his PhD

in Chemistry at the University of Melbourne in 2000, carried out post doctoral research at

Purdue University from 2000–2002 and was an Assistant Member of the Ludwig Institute

for Cancer Research in Melbourne, Australia from 2002–2004. His research interests

involve fundamental and applied studies toward the development of mass spectrometry

strategies for ‘targeted’ proteome and lipidome analysis.

Merlin L. Bruening is a Professor of Chemistry at Michigan State University. Prior to

joining the faculty at Michigan State in 1997, he was an NIH-sponsored postdoctoral

researcher at Texas A&M University. He received his PhD in 1995 from The Weizmann

Institute of Science. His research interests lie in the development of thin films and

membranes for separations, catalysis, and analysis, including sample purification prior to

analysis by mass spectrometry.



Techniques for Phosphopeptide Enrichment Prior to Mass Spec Anal

Documents

strong cation

titanium dioxide

spinach stroma

michigan state

based stationary

spectrometer

tandem mass

carboxylic