Journal of Chromatography A, 1192 (2008) 95–102

Contents lists available at ScienceDirect

Journal of Chromatography A

journa l homepage: www.e lsev ier .com/ locate /chroma

Preparation of a TiO2 nanoparticle-deposited capillary column by liquid phasedeposition and its application in phosphopeptide analysis

Lin

n 430

des fet hi

ten nxidesitioncts.-tubeine wnditi

s pho

Bo Lina,1, Ting Lia,1, Yong Zhaob, Fang-Ke Huangb,a Department of Chemistry, Wuhan University, Wuhan 430072, Chinab College of Life Sciences and State Key Laboratory of Virology, Wuhan University, Wuha

a r t i c l e i n f o

Article history:Received 6 January 2008Received in revised form 7 March 2008Accepted 12 March 2008Available online 20 March 2008

Keywords:Liquid phase depositionNanoparticlesTitanium dioxidePhosphopeptideIn-tube solid phase microextraction

a b s t r a c t

Analysis of phosphopeptisamples is a challenging yamounts, enrichment is ofa thin layer of titanium dioumn by liquid phase depofrom protein digest produdeposition to construct inthe device off-line or on-lof loading and washing cocapillary columns toward

1. Introduction

Protein phosphorylation is an important protein post-translational modification (PTM) and plays a crucial role on variousessential cell functions including signal transduction, metabolicmaintenance and cell division [1–3]. In order to understand howthese cellular processes are controlled by protein phosphorylation,it is essential to determine the sites of protein phosphorylation.Current mass spectrometry (MS)-based phosphorylation siteanalysis strategies [4–9], although successful in many cases, stillpossess difficulty. Due to the low abundance and low ionizationefficiency nature of phosphopeptides, a phosphopeptide enrich-ment sample preparation process is a prerequisite for successfulMS analysis. To be a good phosphopeptide enrichment technique,the method needs not only to be highly selective, but also to beapplicable to samples with limited quantity. In other words, themethod needs to be easy to perform, with high recovery rate, andcost effective.

Commonly used phosphopeptide enrichment methods can beseparated into two main categories based on their enrichmentprinciples: chemical derivatization-based methods and affin-ity purification-based methods [10–17]. Chemical derivatization

∗ Corresponding author. Tel: +86 27 68753800; fax: +86 27 68753797.∗∗ Corresponding author. Tel.: +86 27 87867564; fax: +86 27 68754067.

E-mail addresses: guol@whu.edu.cn (L. Guo), yqfeng@whu.edu.cn (Y.-Q. Feng).1 These authors contributed equally to this work.

0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved.doi:10.1016/j.chroma.2008.03.043

Guob,∗, Yu-Qi Fenga,∗∗

072, China

rom complex mixtures derived from proteolytic digestion of biologicalghly important task. Since phosphopeptides are usually present in smallecessary prior to their characterization by mass spectrometry. In this study,(TiO2) nanoparticles (NPs) was deposited onto the surface of capillary col-(LPD) technique and applied to selectively concentrate phosphopeptides

This is, to our knowledge, the first demonstration of using liquid phasesolid phase microextraction devices for biological analysis. By coupling

ith mass spectrometry analysis, experiments for systematic optimizationons were carried out, and good trapping selectivity of TiO2 NP-depositedsphopeptides was demonstrated.

© 2008 Elsevier B.V. All rights reserved.

methods generally involve multiple chemical reaction steps, andare difficult to apply to complex yet limited protein samples dueto the concerns of sample loss [18–21]. Affinity-based enrichmentmethods include antibody-based affinity chromatography, strongcation and/or anion exchange chromatography, immobilized metalion affinity chromatography (IMAC), metal oxide chromatography,etc. Affinity-based enrichment techniques are the most widely

adopted methods at the moment.

In recent years, a titanium dioxide (TiO2)-based method hasbeen developed for enriching phosphopeptides from complex sam-ples [22–27]. Analytes possessing phosphate functional groups canself-assemble onto the surface of TiO2 particles and elution ofthe bound phosphopeptides can be achieved easily at an alka-line pH. Good compatibility to reversed-phase chromatography andless non-specific binding turn this method into a very interestingalternative to IMAC-based procedures. However, many challengesremain for on-line use of TiO2-based materials with MS detec-tion. Microsized TiO2 particles (5-�m Titanspheres) are commonlyused to pack capillary columns to realize on-line coupling withESI-MS, but the packing and frit preparation are difficult and trou-blesome for many laboratories [23,27]. Nanosized TiO2 particleswere reported to have higher specific surface area and, hence,potentially higher trapping capacities toward phosphopeptidescompared with microsized TiO2 particles [28,29]. However, on-line application of TiO2 nanoparticles with MS has not yet beenrealized due to the lack of appropriate frits to sustain these mate-rials in chromatographic columns, and therefore limited their usein handling a relatively large amount of proteins. Furthermore, the

togr. A

96 B. Lin et al. / J. Chroma

reported preparation of nanoparticles is usually complex and time-consuming.

First reported by Kawahara et al. in 1980s [30,31], liquid phasedeposition (LPD) method is a low-cost, environment-friendly pro-cess for thin film preparation. LPD refers to the formation of oxidethin films from an aqueous solution of a metal-fluoro complexwhich is slowly hydrolyzed by adding water, boric acid (H3BO3) oraluminum metal. The addition of water directly forces precipitationof the oxide, and boric acid (H3BO3) or aluminum metal acts as afluoride scavenger, destabilizing the fluoro complex and promotingprecipitation of the oxide. LPD was first developed for depositingSiO2 thin films [32], and was later used to prepare other metal oxidefilms, such as TiO2, tin oxide, zirconia or a variety of 3d transitionmetal oxides (V, Cr, Mn, Fe, Co, Ni, Cu, Zn, In, individually or com-bined) [33–38]. The chemical and physical properties of obtainedmetal oxide films can be tuned by several experimental parameters,such as pH and concentration for the precursor solution, depositiontime and calcination temperature. LPD is very simple to performand can be easily applied to various kinds of substrates with largesurface area or complex morphology.

In the present work, we deposited a thin TiO2 layer onto theinner surface of capillary column using the mixture of (NH4)2TiF6and boric acid by LPD method. The physical effects of coating cycles,capillary diameter and calcination on the nanostructure of obtainedfilms were characterized by scanning electron microscopy (SEM), X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS).This LPD-based approach is easier to perform and much morecost effective than the previously reported TiO2 coating methodbased on sol–gel process [39–42]. After liquid deposition of TiO2films in capillaries, we examined the feasibility of using such cap-illaries as in-tube solid phase miroextraction (SPME) devices toselectively enrich phosphopeptides from proteolytic digest prior toMS detection. Using LPD technique for the construction of in-tubeSPME devices that are compatible for on-line LC–ESI-MS usage, thisshould open up a new avenue for enriching and analyzing complexbiological samples.

2. Experimental

2.1. Chemicals and materials

Fused-silica capillaries (365 �m o.d., 50 �m i.d. or 365 �m o.d.,75 �m i.d.) were purchased from Yongnian Optic Fiber Plant (Hebei,

China). HPLC grade acetonitrile was obtained from Fisher Scien-tific (USA). Ammonium hexfluorotitanate ((NH4)2TiF6), boric acid(H3BO3) and other chemicals were of analytical reagent grade andsupplied by Shanghai General Chemical Reagent Factory (Shang-hai, China). �-Casein, �-casein and bovine serum albumin (BSA)were purchased from Sigma (St. Louis, MO, USA). A syntheticpeptide from tryptic EGFR with phosphorylation site at aminoacid residue Y1173 (GSTAENAEY(1173)LR, MW 1290), and the cor-responding 15N-labeled peptide (MW 1294) that incorporated15N-Gly, 15N-Ala and 15N-Leu were kindly provided by L.H. Ericsson(Seattle, Washington, USA) [47]. Trypsin was of sequencing gradeand obtained from Promega (Madison, WI, USA). Purified waterwas obtained with a Milli-Q apparatus (Millipore, Bedford, MA,USA).

2.2. Preparation and characterization of TiO2 NP-depositedcapillary columns

Fused-silica capillaries were activated by 1 mol/L NaOH and then1 mol/L HCl. After rinsing with double distilled water, they weredried at 160 ◦C under N2 flow for 10 h.

1192 (2008) 95–102

Mixture of equal volumes of 0.2 mol/L (NH4)2TiF6 and 0.6 mol/LH3BO3 was stirred and used as the precursor solution. The activatedcapillary was filled with the precursor solution, sealed at both endswith silicone rubber and then incubated in a thermostat water bathcontrolled at 35 ± 1 ◦C for 16 h for LPD. After deposition, the capil-lary was washed with distilled water and dried at 120 ◦C for 4 hunder constant N2 flow. Fluorin residues in the TiO2 film from theprecursor solution were removed by washing the capillary with0.1 mol/L NaOH and distilled water in sequence. The calcination ofTiO2 NP-deposited capillary column was performed by heating ata rate of 1 ◦C/min to 300 ◦C and holding for 2 h to age the TiO2 filmand increase cross-linking of the inorganic framework.

The morphology of the capillary inner surface was displayedby QUANTA-200 SEM (FEI, The Netherlands). In order to furthercharacterize the TiO2 film, LPD was also carried out on a quartzplate in a similar way to that performed in capillaries. The existenceof TiO2 film and its composition were determined by XSAM800X-ray photoelectron spectroscopy (XPS, Kratos, UK), with Mg K�radiation as the exciting source. The crystal structure of TiO2 filmwas determined with a XRD-6000 X-ray diffractometer (Shimadzu,Japan) using Cu K� radiation and a rotating anode operated at 40 kVand 30 mA.

2.3. Sample preparation

�-Casein and �-casein were originally made up into stock solu-tions of 1 mg/mL using distilled water. Proteins were digested usingtrypsin at an enzyme to substrate ratio of 1:50 (w/w) in 2 mol/Lurea, 100 mmol/L Tris pH 8.5 overnight at 37 ◦C; the digest mixtureswere then stored at −20 ◦C without further treatment.

BSA (4 mg) was dissolved in 1 mL denaturing buffer solutioncontaining 8 mol/L urea in 50 mmol/L ammonium bicarbonate. Theobtained protein solution was mixed with 20 �L of 50 mmol/Ldithiothreitol (DTT) and incubated for 15 min at 65 ◦C to reduceprotein disulfide bonding. 40 �L of 50 mmol/L iodoacetamide (IAA)was added, and the obtained solution was incubated for an addi-tional 30 min at room temperature in dark. The above reducedand alkylated protein mixture was diluted 10-fold with 50 mmol/Lammonium bicarbonate and incubated overnight at 37 ◦C withtrypsin at an enzyme to substrate ratio of 1:50 (w/w) to produceproteolytic digest mixture.

In order to control the pH value and ionic strength moreprecisely during the optimization of loading/washing/eluting con-ditions, we chose to use desalted tryptic digest mixture as

experimental material. Desalting was achieved using C18 Zip-Tip pipette tips (Millipore, Billerica, MA) and the bound peptideswere eluted by 50% acetonitrile/0.1% trifluoroacetic acid solution.Desalted digest mixtures were dried by speed-vac centrifugation,and were subsequently brought up in different concentrations ofacetonitrile and acetic acid solution as specified in the text.

2.4. Enrichment of phosphorylated peptides using TiO2NP-deposited capillary columns

The extraction of tryptic digest was performed using a sim-ple laboratory-made system [46], which is a combination of twomicroflow pumps (pump A and B), a six-port valve switching mod-ule (Unimicro Technologies, Shanghai, China) and a stainless steelsample loop (50 �L) (see Fig. 1). The peptides were first loaded ontothe column in the solution of 0.1% (v/v) acetic acid/acetonitrile (1:9,v/v) at a flow rate of 20 �L/min by pump A. After washing withthe same solution for 2.5 min at 20 �L/min, the six-port valve Bwas switched and the bound peptides were eluted with the mix-ture of 0.3% (v/v) ammonium hydroxide/acetonitrile (5:5, v/v) at20 �L/min for 0.5 min by pump B.

B. Lin et al. / J. Chromatogr. A

Fig. 1. Construction of in-tube SPME/ESI-Q-TOF MS.

2.5. Mass spectrometry and data analysis

MALDI-TOF spectra of peptides were obtained using a Voy-ager DE STR MALDI-TOF work station mass spectrometer (AppliedBiosystems Inc., USA). The analysis was performed in positive ionreflector mode with an accelerating voltage of 20 kV and a delayedextraction of 280 ns. Typically, 200 scans were averaged. For theanalysis of phosphopeptides, 20 mg/mL 2,5-dihydroxybenzoic acid(DHB) in 50% acetonitrile, 1% phosphoric acid was used as matrixsolution. 0.5 �L of sample was mixed with matrix solution at 1:1ratio prior to be deposited onto the target plate. The amount ofpeptides spotted onto target was estimated to be around 400 fmol.

Negative ion ESI mass spectra were acquired on a Bruker (Bre-men, Germany) Q-TOF instrument. The instrument was tunedand calibrated according to manufacturer’s specifications beforeuse. For optimization of ionization, desalted tryptic digest of �-casein was diluted to 0.1 pmol/�L with 0.3% (v/v) ammoniumhydroxide/50% acetonitrile before direct infusion into the mass

Fig. 2. Setup of in-tube

1192 (2008) 95–102 97

spectrometer at a flow rate of 20 �L/min by a syringe pump (KdScientific Inc., Holliston, MA, USA). The optimal ionization sourceworking parameters were listed as follows: capillary voltage 4.5 kV;ion energy of quadruple −5 eV/z; dry temperature 180 ◦C; nebulizer0.6 bar, dry gas 5.0 L/min. Instrument control and data analysis wereperformed using Bruker Daltonics Data analysis 3.4 software.

On-line coupling of the TiO2 NP-deposited capillary columnwith reversed-phase liquid chromatography (LC)–ESI-MS/MS wasperformed on a Applied Biosystems Qtrap 3200 instrument (AB,Foster City, CA, USA) equipped with a Tempo nano MDLC sys-tem (Eksigent, USA) and a nanoelectrospray source (see Fig. 2).Peptide mixtures were loaded onto a doubly TiO2 NP-depositedcapillary in 20% acetonitrile solution using an autosampler, washedwith acetic acid (0.1%, v/v) at 1 �L/min for 25 min, and eluted with0.3% (v/v) ammonium hydroxide at 500 nL/min for 25 min directlyonto a homemade 10–15 cm analytical C18 reversed-phase column(100 �m i.d.; 360 �m o.d.; YMC ODS-AQ 5 �m 200 A (Waters, Mil-ford, MA, USA)). Peptides from the reversed-phase column wereeluted using a linear gradient of 0–100% solution B (solution A:2% ACN, 0.1% formic acid; solution B: 98% ACN, 0.1% formic acid)in 75 min. The C18 column was connected to a coated fused silicaemitter (75 �m i.d.; 360 �m o.d.; 15 �m tip i.d.) (New Objective,

Cambridge, MA) through a fused silica transfer line (20 �m i.d.).Data-dependent MS/MS mode was used, and peptides were identi-fied by searching a non-redundant protein sequence database usingthe Mascot program [43] and the ProteinPilotTM 2.0.1 software (AB,Foster City, CA, USA).

3. Results and discussion

3.1. Preparation and characterization of TiO2 NP-depositedcapillary columns

The formation of TiO2 film by LPD can be represented as follows:

TiF62− + nH2O ⇔ [TiF(6−n)OHn]2− + nHF (1)

[TiF(6−n)(OH)n]2−OH−−→[TiF(5−n)(OH)(n+1)]

2− + F−

(5−n)OH−−→ [Ti(OH)6]2− + (6 − n)F− (2)

SPME-LC–MS/MS.

98 B. Lin et al. / J. Chromatogr. A 1192 (2008) 95–102

thoutfter tw

Fig. 3. Scanning electron microscopy images of inner surfaces of capillaries: (a) wi50 �m; (c) after one deposition in the capillary with inner diameter of 75 �m; (c) a

H3BO3 + 4HF ⇔ BF4− + H3O+ + 2H2O (3)

As shown in Eqs. (1) and (2), hydrolysis of TiF62− takes place

by a ligand-exchange reaction. The equilibrium will be shiftedtowards the oxide if the concentration of water is increased orif the hydrogen fluoride concentration is decreased. Boric acid

reacts readily with the product HF to form the more stable BF4

−

ions, which promotes the consumption of non-coordinated F−

ions and the production of water (Eq. (3)) to shift reaction 1 tothe right-hand side. Homogeneous nucleation and colloid forma-tion would occur once the oxide’s solubility limit is reached, andthin films are formed via attraction and attachment of colloidalparticles. The films were reported to form preferentially on SiO2-covered substrates with reactive hydroxyl groups, and chemicalbonds of Ti–O–Ti and Ti–O–Si were believed to form during thedeposition for strong adhesion with the substrates [45]. Here wetook full advantage of the experiences obtained by others in thepreparation of TiO2 films and applied them to fused-silica capil-lary columns. The preparation of the precursor solution as wellas the molar ratio of (NH4)2TiF6 and H3BO3 we used was simi-lar to that reported by Feng et al. [44], and our investigation wasfocused on the influence of capillary diameter, coating cycle andcalcination on the microstructure of TiO2 films within capillarycolumns.

Fig. 3(b) and (c) shows SEM micrographs of the TiO2 thin filmsdeposited on the surface of capillary columns with inner diam-eters of 50 and 75 �m, respectively. It can be observed that the

deposition of TiO2; (b) after one deposition in the capillary with inner diameter ofo depositions in the capillary with inner diameter of 50 �m.

surface morphologies and roughness of the TiO2 films are almostthe same in two different capillaries. In spite of some cracks, the filmis overall dense and uniform, which is composed of un-sphericalnanoparticles with a narrow diameter range from 100 to 200 nm.The un-spherical nanoparticles are supposed to be the aggregatesof many smaller TiO2 particles, and the cracks could be ascribed to

the increase of internal stress from the shrinkage of thin film. Afterthe TiO2 NP-deposited capillary was heated at 120 ◦C for 4 h undernitrogen atmosphere and treated with another round of deposi-tion (Fig. 3(d)), dramatic changes occur in the film morphology.The particle size is much bigger and the shape of the particles ismore irregular than that obtained with one deposition, which couldbe caused by either the growth of initially formed particles or theaggregation of newly formed particles.

The calcination after liquid phase deposition is considered tofavor the crystallization of TiO2 particles. Improvement of TiO2crystallization after calcination was proved by XRD experiments(Fig. 4). Since it was difficult to acquire XRD spectrum of TiO2 filmdirectly from the TiO2 NP-deposited capillary, a fused quartz sub-strate was treated by the same procedure of LPD as that in thecapillary. The XRD results of TiO2 film on the fused quartz indicatethat the TiO2 film calcinated at 300 ◦C shows preferential orienta-tion for anatase in the (1 0 1) direction. The average crystallite sizeof the film calculated using Debye–Scherrer formula for the maindiffraction peak is determined to be about 8.6 nm.

The composition of TiO2 film was characterized by XPS, whichwas also carried out on the flat quartz plate (Fig. 5(a)). Results show

B. Lin et al. / J. Chromatogr. A

Fig. 4. X-ray diffraction patterns of TiO2 films deposited on the quartz plate beforecalcination (a) and after calcination at 300 ◦C (b).

that the film contains not only Ti and O elements, but also a smallamount of F element which is identified to be from the precursor

solution. The existence of F is expected to weaken the interaction ofphosphate groups with TiO2 materials in later trapping of phospho-peptides from protein digest products, but rinsing with 0.1 mol/LNaOH can completely eliminate the F element (Fig. 5(b)).

3.2. Evaluation of the phospho-selectivity of TiO2 NP-depositedcapillary columns by MALDI-MS

After preparation of the TiO2 NP-deposited capillary, we testedwhether this capillary column can be used as a novel in-tube SPMEdevice in the preconcentration and isolation of phosphopeptidesfrom proteolytic digests. The TiO2 NP-deposited capillary column(570 mm × 50 �m i.d.) without calcination was used in the follow-ing extractions unless otherwise stated. The capillary was installedonto a six-port valve, and loading and elution solutions were drivenby two mircoflow pumps to achieve the selective isolation of phos-phopeptides from protein digests. After loading 50 �L of digestedsample (0.08 pmol/�L) onto the TiO2 NP-deposited capillary col-umn in a loading solution (e.g. 0.1% (v/v) acetic acid) at 20 �L/min,the column was washed with the same loading solution for 2.5 min

Fig. 5. X-ray photoelectron spectroscopy spectra of TiO2 films deposited on thequartz plate before NaOH rinsing (a) and after NaOH rinsing (b).

1192 (2008) 95–102 99

at 20 �L/min. Subsequently, the trapped peptides were eluted with10 �L of an elution solution (e.g. 0.3% (v/v) ammonium hydroxide)at 20 �L/min, and collected for MS analysis by MALDI-TOF.

The enrichment of phosphopeptides from tryptic digests of �-casein and �-casein were tested, respectively, using the abovedescribed procedure. After enrichment, all non-phosphorylatedpeptides from protein digests were either not detectable orproduced signals with drastically reduced intensities comparedwith non-enriched samples. For tryptic digest of �-casein, twophosphopeptides, FQ[pS]EEQQQTEDELQDK (�1, m/z 2062.8) andRELEELNVPGEIVE[pS]L[pS][pS][pS]EESITR (�2, m/z 3122.3), couldbe unambiguously detected. For tryptic digest of �-casein, eightunique phosphopeptides were identified after enrichment. Com-plete information for the enriched phosphopeptides from �-caseindigest, including amino acid sequence and phosphorylation sites, isshown in Table 1. The above results demonstrated the effectivenessof the TiO2 film as a phosphopeptide enrichment material.

During the preparation of TiO2 NP-deposited capillary columns,the calcination temperatures may affect the crystallization of TiO2films and thus change their binding specificity. Therefore we inves-tigated the influence of calcination on the extraction performanceof TiO2 films by carrying out parallel experiments between cal-cinated and non-calcinated capillary columns. Results show thephosphopeptide enrichment is similar between films with or with-out heat treatment. To simplify the preparation procedure, TiO2NP-deposited capillary columns without calcination were used forall the experiments described below to explore the full potentialuse of such devices on-line coupled to ESI-MS.

3.3. On-line coupling of TiO2 NP-deposited capillary columnswith ESI-MS

One of the major advantages for the in-tube SPME tech-nique over other phosphopeptide enrichment strategies is its goodcompatibility with subsequent analysis procedures, especially itseasiness to be coupled with on-line ESI-MS analysis. In order tospeed up the optimization processes, we bypassed the LC sepa-ration and connected the extraction process directly with ESI-MSdetection. Trypsin digest of �-casein was used as the standard sam-ple.

Through repeated testing, we found 50% acetonitrile with 0.3%(v/v) ammonium hydroxide was a good elution solution. Theaddition of acetonitrile not only enhanced elution by reducinghydrophobic interactions of peptides with the TiO2 film, but also

produced better ESI spray for the ionization of phosphopeptides.With the above basic elution solution, negative ion detection modewas used and most of phosphopeptides were detected as deproto-nated ions.

The loading and washing conditions were then optimized forphosphopeptide enrichment. Acidic condition (pH 2–3) is usuallyemployed in the enrichment of phosphopeptides using TiO2 tomake sure that the acidic residues in the peptides are neutral. Sixsingly phosphorylated peptides and one doubly phosphorylatedpeptide (Table 1) were identified using 0.1% (v/v) acetic acid (pH3) as the loading and washing solutions. In order to improve theextraction selectivity, acetonitrile was added to reduce the absorp-tion of non-phosphopeptides onto the TiO2 film. The acetonitrilecontent was investigated in the range of 0–90%. Finally, 90% ace-tonitrile in 0.1% (v/v) acetic acid was used as loading and washingsolutions to achieve best MS detection results.

According to our investigation, only four singly phosphorylatedpeptides (m/z 731.8, 828.9, 922.3 and 974.5) and one doubly phos-phorylated peptide (m/z 962.4) could be observed by direct MSanalysis. After extraction of tryptic digest under optimal condi-tions, two additional singly phosphorylated peptides (m/z 795.8

togr. A

c dige

e

RRDIKMEDI

100 B. Lin et al. / J. Chroma

Table 1Overview of the phosphopeptides identified by different MS techniques from trypti

No. [M+H]+ Phosphorylation site Amino acid sequenc

�1 1466.6 1 TVDME[pS]TEVFTK�2 1594.7 1 TVDME[pS]TEVFTKK�3 1660.8 1 VPQLEIVPN[pS]AEE�4 1832.8 1 YLGEYLIVPN[pS]AEE�5 1847.6 1 DIG[pS]ESTEDQAME�6 1927.7 2 DIG[pS]E[pS]TEDQA

�7 1952.0 1 YKVPQLEIVPN[pS]AEER�8 2678.0 3 VNEL[pS]KDIG[pS]E[pS]TE�9 2935.1 3 KEKVNEL[pS]KDIG[pS]E[p

and 914.9) are detected following enrichment (Table 1). Comparedto the results obtained by MALDI-TOF MS, ESI-QTOF MS analy-sis was more biased towards singly phosphorylated peptides. Thedetection sensitivity of phosphopeptide by this ESI-MS method wasestimated by using peptide amount ranging from 0.1 to 10 pmol. Allseven phosphopeptides can be detected when the peptide amountwas higher than 0.5 pmol; at lower peptide level, the detection ofsome phosphopeptides may be reduced or missing. In addition,since phosphopeptides could still be detected in the sample flow-through, the binding efficiency of TiO2 materials was not 100%,which implied the SPME mechanism.

Before the TiO2 NP-deposited capillary was applied to morecomplex mixtures, we investigated the influence of capillary diam-eter on the enrichment of phosphopeptides. TiO2 thin films weredeposited onto the surface of capillaries with inner diameters of 50

Fig. 6. Negative ion ESI-QTOF mass spectra for the digest mixtures of �-casein (2 pmol) acolumns prepared by one deposition (a) and two deposition (b), respectively. All the iden

1192 (2008) 95–102

st of �-casein after enrichment with TiO2 NP-deposited capillary column

MALDI-MS ESI-MS LC–ESI-MS/MS

+ + ++ + ++ + ++ + −− + −

K + + +

+ + +

DQAMEDIK + − −S]TEDQAMEDIKQ + − −

and 75 �m, respectively. Parallel extractions using these two cap-illary columns at the same linear velocity of mobile phase werecarried out. The S/N ratios of phosphopeptides are much higher forthe 50 �m capillary column. This can be ascribed to more efficientcapture of phosphopeptides with 50 �m TiO2 NP-deposited capil-lary due to its higher ratio of solid phase to mobile phase. Althoughcapillaries with smaller inner diameters were favorable for effi-cient extractions, we did not further downsize the capillary we usedsince capillary with the inner diameter below 50 �m was difficultto operate and prepare due to high back pressure.

To evaluate the ability to capture the phosphopeptides from acomplex sample, the TiO2 NP-deposited capillary was applied totrap phosphopeptides from the mixtures of tryptic digest of �-casein and BSA (non-phosphoprotein). Fig. 6(a) presents the ESI-MSspectra of the sample after using TiO2 NP-deposited capillary to

nd BSA at different molar ratios after enrichment with TiO2 NP-deposited capillarytified phosphopeptides are labeled with asterisks.

B. Lin et al. / J. Chromatogr. A 1192 (2008) 95–102 101

otal ioTable

Fig. 7. Positive ion ESI-Qtrap mass spectra of �-casein tryptic digest (5 pmol). (a) Tin-tube SPME-LC–MS/MS. The identified phosphopeptides are labeled and listed inand 976.9.

enrich phosphopeptides from the tryptic digest of �-casein andBSA with molar ratio of 1:0, 1:1 and 1:10, respectively. Similarto the results from pure �-casein digest, seven phosphopeptidescan be detected in the mass spectrum, but ion signals were obvi-ously reduced when BSA was mixed with �-casein at molar ratioof 1:1. When the amount of BSA was further increased, phos-phopeptides from �-casein were difficult to distinguish amongmany abundant non-phosphopeptides peaks. A real biological sam-ple was expected to be more complex than our tested mixtures,

and in our effort to find a better procedure for handling complexsamples, we went back to the LPD protocol. TiO2 NP-depositedcapillary was treated with another round of deposition, and thephosphopeptide enrichment results were shown in Fig. 6(b). For thedoubly deposited capillary, phosphopeptide signals were greatlyenhanced compared to that obtained by the singly deposited cap-illary under same extraction conditions. We considered that theimproved phospho-specificity of doubly deposited capillary wasdue to the better coverage of Si–OH on the inner capillary sur-face which gave rise to non-specific binding. The optimal loadingcondition for the doubly deposited capillary was the same as thesingle deposited capillary, but for washing solution 5% (v/v) aceticacid solution was used to reduce non-phosphopeptide signals. Itis worth mentioning that in the former optimization for the singlydeposited capillary, no phosphopeptides were observed when 4%(v/v) acetic acid was used for washing. It was probably that onemore deposition process increased the amount of TiO2 material andhence enhanced its ability of trapping phosphopeptides, so higherconcentration of acetic acid could be used for the doubly depositedcapillary. As shown in Fig. 6(b), phosphopeptides can be easilydetected from the digest mixtures of �-casein and BSA with molar

n chromatogram obtained by LC–MS/MS; (b) total ion chromatogram obtained by1; (c, d) selected ion monitoring spectra of two phosphopeptides with m/z 798.7

ratio of 1:1 and 1:10 after enrichment with the doubly depositedcapillary. The recovery yield of the doubly deposited capillarywas evaluated for a synthetic phosphorylated peptide from trypticEGFR (GSTAENAEY(1173)LR, MW 1290), using the corresponding15N-labeled peptide (MW 1294) as the internal standard (IS) toassure the same ionization efficiency [47]. By comparing analyte/ISratios in mass spectra of the standard sample solution (20 pmol)before and after enrichment, the recovery was determined to be67%.

3.4. On-line coupling of TiO2 NP-deposited capillary columnswith LC–ESI-MS/MS

For complex peptide mixture analysis, reversed-phase LC is themost commonly used separation technique prior to the ESI-MSanalysis. Therefore, doubly TiO2 NP-deposited capillary was usedto test whether such an in-tube SPME device can be effectivelylinked to a standard nanoLC–ESI-MS/MS system using a Qtrap3200instrument from Applied Biosystems. Total ion chromatograms fortryptic digest of �-casein obtained before and after enrichment areshown in Fig. 7. Compared to the starting �-casein peptide mixture(Fig. 7(a)), the detection of phosphopeptides is greatly improvedafter pretreatment using our TiO2 NP-deposited capillary columnbefore LC–MS/MS analysis (Fig. 7(b)–(d)). Five phosphopeptideswere detected when 5 pmol of �-casein digest was used (Table 1);when the �-casein level was lowered to 50 fmol, we could stilldetect two phosphopeptides with m/z 831.2 and 976.9. The digestmixture of 500 fmol �-casein and 50 pmol BSA was also tested byin-tube SPME-LC–MS/MS. Results show five phosphopeptides from�-casein can still be easily identified, which further reveals the

togr. A

[

[[

[[

[[

[

[[

[[[[

[

[

[26] A. Schlosser, J.T. Vanselow, A. Kramer, Anal. Chem. 77 (2005) 5243.[27] G.T. Cantin, T.R. Shock, S.K. Park, H.D. Madhani, J.R. Yates, Anal. Chem. 79 (2007)

102 B. Lin et al. / J. Chroma

effectiveness of TiO2 NP-deposited capillary column as an on-lineenriching device for phosphopeptide analysis.

4. Conclusions

Using LPD technique, a novel in-tube SPME device, a thinTiO2 film within fused-silica capillary column, was constructedand applied off-line or on-line with MS detection to isolate andanalyze phosphopeptides from tryptic digest of phosphoproteinsfor the first time. The preparation for these TiO2 NP-depositedcapillary columns is simple and the cost is low. Good trappingcapability for phosphopeptides was proved by both off-line usewith MALDI-TOF MS and on-line use with ESI-QTOF MS. Auto-mated enrichment and analysis of phosphopeptides can be easilyaccomplished by coupling this in-tube SPME device on-line with astandard nanoLC–ESI-MS/MS. The detection limit for phosphopep-tide was determined to be 50 fmol or lower for in-solution trypticdigest of �-casein under optimal conditions, and good enrichmentwere obtained in the analysis of tryptic digest mixtures of �-caseinand BSA. During on-line investigation of TiO2 NP-deposited cap-illary columns coupled to ESI-MS, TiO2 films were stable afterhundreds of extractions and caused no detachment problems forMS detection. Our research provided a new and low cost strategyfor on-line coupling of sample pretreatment with MS detection for

the analysis of phosphopeptides.

Acknowledgements

This work was partly supported by grants from the NationalNatural Science Foundation of China (Grant No. 20475040),the National Science Fund for Distinguished Young Scholars(No. 20625516), the Science Fund for Creative Research Groups(No. 20621502), NSFC and the National 973 Project of China(2006CB504300 and 2007CB914200).

References

[1] L.N. Johnson, R.J. Lewis, Chem. Rev. 101 (2001) 2209.[2] M. Bollen, M. Beullens, Trends Cell Biol. 12 (2002) 138.[3] R.E. Honkanen, T. Golden, Curr. Med. Chem. 9 (2002) 2055.[4] D.T. McLachlin, B.T. Chait, Curr. Opin. Chem. Biol. 5 (2001) 591.[5] M. Mann, S.E. Ong, M. Gronborg, H. Steen, O.N. Jensen, A. Pandey, Trends

Biotechnol. 20 (2002) 261.[6] R.E. Schweppe, C.E. Haydon, T.A. Lewis, K.A. Resing, N.G. Ahn, Acc. Chem. Res.

36 (2003) 453.[7] M.J. Chalmers, W. Kolch, M.R. Emmett, A.G. Marshall, J. Chromatogr. B 803

(2004) 111.

[[[

[

[[[

[[[[[[[[[

[[[[

1192 (2008) 95–102

[8] G.T. Cantin, J.R. Yates, J. Chromatogr. A 1053 (2004) 7.[9] F. Meng, A.J. Forbes, L.M. Miller, N.L. Kelleher, Mass Spectrom. Rev. 24 (2005)

57.10] C.S. Raska, C.E. Parker, A. Dominski, W.F. Marzluff, G.L. Glish, R.M. Pope, C.H.

Borchers, Anal. Chem. 74 (2002) 3429.11] A. Stensballe, S. Andersen, O.N. Jensen, Proteomics 1 (2001) 207.12] T.S. Nuhse, A. Stensballe, O.N. Jensen, S.C. Peck, Mol. Cell. Proteomics 2 (2003)

1234.13] M.C. Posewitz, P. Tempst, Anal. Chem. 71 (1999) 2883.14] S.B. Ficarro, M.L. McCleland, P.T. Stukenberg, D.J. Burke, M.M. Ross, J. Sha-

banowitz, D.F. Hunt, F.M. White, Nat. Biotechnol. 20 (2002) 301.15] K. Moser, F.M. White, J. Proteome Res. 5 (2006) 98.16] A. Pandey, A.V. Podtelejnikov, B. Blagoev, X.R. Bustelo, M. Mann, H.F. Lodish,

Proc. Natl. Acad. Sci. U.S.A. 97 (2000) 179.17] S.B. Ficarro, O. Chertihin, V.A. Westbrook, F. White, F. Jayes, P. Kalab, J.A. Marto,

J. Shabanowitz, J.C. Herr, D.F. Hunt, P.E. Visconti, J. Biol. Chem. 278 (2003)11579.

18] H. Zhou, J.D. Watts, R. Aebersold, Nat. Biotechnol. 19 (2001) 375.19] Z.A. Knight, B. Schilling, R.H. Row, D.M. Kenski, B.W. Gibson, Nat. Biotechnol. 21

(2003) 1047.20] M.P. Molloy, P.C. Andrews, Anal. Chem. 73 (2001) 5387.21] D.T. McLachlin, B.T. Chait, Anal. Chem. 75 (2003) 6826.22] A. Sano, H. Nakamura, Anal. Sci. 20 (2004) 861.23] M.W.H. Pinkse, P.M. Uitto, M.J. Hilhorst, B. Ooms, A.J.R. Heck, Anal. Chem. 76

(2004) 3935.24] M.R. Larsen, T.E. Thingholm, O.N. Jensen, P. Roepstorff, T.J.D. Jorgensen, Mol.

Cell. Proteomics 4 (2005) 873.25] C. Klemm, S. Otto, C. Wolf, R.F. Haseloff, M. Beyermann, E. Krause, J. Mass Spec-

trom. 41 (2006) 1623.

4666.28] C.-T. Chen, Y.-C. Chen, Anal. Chem. 77 (2005) 5912.29] S.S. Liang, H. Makamba, S.Y. Huang, S.H. Chen, J. Chromatogr. A 1116 (2006) 38.30] H. Kawahara, H. Honda, Japanese Patent 59141441 A (Nippon Sheet Glass)

(August 14, 1984).31] J. Ino, A. Hishinuma, H. Nagayama, H. Kawahara, Japanese Patent 01093443 A

(Nippon Sheet Glass) (April 12, 1989).32] H. Nagayama, H. Honda, H. Kawahara, J. Electrochem. Soc. 135 (1988) 2013.33] S. Deki, Y. Aoi, O. Hiroi, A. Kajinami, Chem. Lett. 25 (1996) 433.34] K. Shimizu, H. Imai, H. Hirashima, K. Tsukuma, Thin Solid Films 351 (1999)

220.35] S. Deki, Y. Aoi, Y. Miyake, A. Gotoh, A. Kajinami, Mater. Res. Bull. 31 (1996) 1399.36] S. Deki, Y. Aoi, A. Kajinami, J. Mater. Sci. 32 (1997) 4249.37] K. Tsukuma, T. Akiyama, H.J. Imai, Non-Cryst. Solids 210 (1997) 48.38] Y.F. Gao, Y. Masuda, T. Yonezawa, K. Koumoto, Chem. Mater. 14 (2002) 5006.39] P. Tsai, C.-T. Wu, C.S. Lee, J. Chromatogr. B 657 (1994) 285.40] C. Fujimoto, Electrophoresis 23 (2002) 2929.41] L. Xu, Y.-Q. Feng, Z.-G. Shi, S.-L. Da, Y.-Y. Ren, J. Chromatogr. A 1028 (2004) 165.42] Y.L. Hsieh, T.H. Chen, C.P. Liu, C.Y. Liu, Electrophoresis 26 (2005) 4089.43] D.N. Perkins, D.J. Pappin, D.M. Creasy, J.S. Cottrell, Electrophoresis 20 (1999)

3551.44] L. Feng, Y.J. Liu, J.M. Hu, Langmuir 20 (2004) 1786.45] M.-K. Lee, B.-H. Lei, Jpn. J. Appl. Phys. 39 (2000) L101.46] B. Lin, M.-M. Zheng, S.-C. Ng, Y.-Q. Feng, Electrophoresis 28 (2007) 2771.47] L. Guo, C.J. Kozlosky, L.H. Ericsson, T.O. Daniel, D.P. Cerretti, R.S. Johnson, J. Am.

Soc. Mass Spectrom. 14 (2003) 1022.