Electrostatic Repulsion-Hydrophilic Interaction Chromatography

Oct 20, 2011 - Hypoxia-induced Changes to Integrin ? 3 Glycosylation Facilitate Invasion in Epidermoid Carcinoma Cell Line A431. Yan Ren , Piliang Hao...
0 downloads 5 Views 2MB Size
TECHNICAL NOTE pubs.acs.org/jpr

Electrostatic Repulsion-Hydrophilic Interaction Chromatography (ERLIC) versus Strong Cation Exchange (SCX) for Fractionation of iTRAQ-Labeled Peptides Piliang Hao,†,‡ Jingru Qian,† Yan Ren,† and Siu Kwan Sze*,†,‡ † ‡

School of Biological Sciences, Nanyang Technological University, 60 Nanyang Drive, 637551 Singapore Singapore Centre on Environmental Life Sciences Engineering, Nanyang Technological University, 60 Nanyang Drive, 637551 Singapore

bS Supporting Information ABSTRACT: The iTRAQ technique is popular for the comparative analysis of proteins in different complex samples. To increase the dynamic range and sensitivity of peptide identification in shotgun proteomics, SCX chromatography is generally used for the fractionation of iTRAQ-labeled peptides before LC MS/MS analysis. However, SCX suffers from clustering of similarly charged peptides and the need to desalt fractions. In this report, SCX is compared with the alternative ERLIC method for fractionating iTRAQ-labeled peptides. The simultaneous effect of electrostatic repulsion and hydrophilic interaction in ERLIC results in peptide elution in order of decreasing pI and GRAVY values (increasing polarity). Volatile solvents can be used. We applied ERLIC to iTRAQ-labeled peptides from rat liver tissue, and 2745 proteins and 30 016 unique peptides were identified with high confidence from three technical replicates. This was 12.9 and 49.4% higher, respectively, than was obtained using SCX. In addition, ERLIC is appreciably better at the identification of highly hydrophobic peptides. The results indicate that ERLIC is a more convenient and more effective alternative to SCX for the fractionation of iTRAQ-labeled peptides. Quantification data show that both SCX and ERLIC fractionation have no significant effect on protein quantification by iTRAQ. KEYWORDS: iTRAQ, ERLIC, SCX, MMC

’ INTRODUCTION iTRAQ (isobaric tagging for relative and absolute quantitation) labeling has become popular for comparative analysis of proteins in complex samples because of its capacity for multiplexing and its accuracy.1 3 The proteins in four or eight biological samples can be labeled and quantified in a single analysis. Theoretically, the primary amines of all tryptic peptides can be labeled so that multiple peptides from the same protein are identified and quantified, facilitating statistical analysis for identification and quantification with high confidence. In contrast to stable isotope labeling with amino acids in cell culture (SILAC), iTRAQ can be applied to tissue, serum, and other complex samples in addition to cultured cells. A total of 555 publications were retrieved with the key word “iTRAQ” in PubMed by July 2011; 281 of them were published in 2010 and 2011, indicating the increasing popularity of iTRAQ. To reduce sample complexity for improving dynamic range and proteome coverage, multidimensional liquid chromatography (MDLC) is generally used in the fractionation of iTRAQlabeled peptides.4 The combination of offline SCX (strong cation exchange) with RP (reverse phase) chromatography is the most widely used MDLC sequence because of the good orthogonality of the methods; peptides are separated on the basis of charge and hydrophobicity, respectively.5 7 However, some clustering of r 2011 American Chemical Society

similarly charged peptides is inevitable in SCX because most tryptic peptides carry two (+) charges at the pH of 3.0.8 In addition, high concentrations of nonvolatile salts are generally used for eluting peptides in SCX. Online desalting of the collected fractions from SCX with the trap column on the HPLC coupled to MS/MS was used in some laboratories,9 but high concentration of nonvolatile salts can frequently lead to the blockage of the autosampler of the HPLC and the quick deterioration of the trap column. Thus, the peptide fractions have to be desalted using RP cartridges before being subjected to LC MS/MS in order to avoid ionization suppression and plugging of the separation capillary or MS orifice. The desalting may result in some degree of peptide loss. It is also expensive and time-consuming because dozens of fractions need to be desalted in a single analysis. Some alternatives to SCX have been evaluated for iTRAQlabeled peptides. For example, the compatibility of iTRAQ with OFFGEL isoelectric focusing with immobilized pI strips was demonstrated in 200710 but has been employed to date in only a few publications.11,12 Recently, the separation of iTRAQ-labeled peptides was evaluated on a mixed-mode column with both RP and anion exchange (AX) properties, and increased proteome Received: August 11, 2011 Published: October 20, 2011 5568

dx.doi.org/10.1021/pr2007686 | J. Proteome Res. 2011, 10, 5568–5574

Journal of Proteome Research coverage and more uniform distribution of peptides in the collected fractions was reported compared with SCX analysis;13 hydrophilic interaction chromatography (HILIC) was reported to minimize iTRAQ ratio compression through high-resolution fractionation of peptides;14 high pH RP was also applied to the fractionation of iTRAQ-labeled peptides with significantly better performance than the traditional one-dimensional SDS-PAGE and OFFGEL isoelectric focusing.15,16 In addition, three-dimensional MDLCs, such as online RP-SCX-RP with SCX as trap columns and offline SCX followed by online RP-RP, were also evaluated for their potential in identifying and quantifying iTRAQ-labeled peptides, and significantly better performances were claimed compared with the traditional methods.17,18 The introduction of a further dimensional separation does improve the overall separation power of the MDLC, but it generally results in the increase of instrument time, and the online MDLC is mostly not as stable as offline ones. The new ERLIC mode of chromatography was first introduced by Alpert for separation of biomolecules and phosphopeptide enrichment.19 It has good resolving power and is orthogonal to RP chromatography; thus, it is suitable for first dimensional fractionation of peptides in 2D LC MS/MS. Salt-free mobile phases have been developed for convenient sample preparation. In our previous study, we introduced an ERLIC method for fractionating peptides evenly from rat kidney tissue using its mixed-mode chromatographic (MMC) properties. This method separates peptides on the basis of the simultaneous effect of electrostatic repulsion and hydrophilic interaction, with elution in order of high to low pI and GRAVY values.20 Here, we extend the application of ERLIC to the fractionation of iTRAQlabeled peptides for improving the 2D LC MS/MS workflow and increasing proteome coverage.

’ MATERIALS AND METHODS Sample Preparation and Digestion

Male Sprague Dawley rats were handled in accordance with the guidelines of NTU Institutional Animal Care and Use Committee (NTU-IACUC), NTU, Singapore. Rat livers were snap-frozen in liquid nitrogen immediately after collection and kept at 80 °C until use. The tissue was ground into fine powder in liquid nitrogen. The powder was then suspended in 8 M urea/ 50 mM triethylammonium bicarbonate (TEAB) with a protease inhibitor cocktail added according to instructions from the manufacturers. The suspension was sonicated for 10 s thrice on ice and centrifuged at 20000g at room temperature (RT) for 20 min. The protein concentration of the supernatant was then determined by the bicinchoninic acid (BCA) assay. Proteins were first reduced in 5 mM Tris-(2-carboxyethyl) phosphine (TCEP) for 1 h at 60 °C, followed by blocking cysteine residues in 10 mM methyl methanethiosulfate (MMTS) for 30 min at RT in the dark. After the concentration of urea was diluted to 1 M with 50 mM TEAB, trypsin was added at a ratio of 1:50 (trypsin/ sample). It was then incubated at 37 °C overnight. The tryptic peptides so obtained were desalted using Sep-Pak C18 cartridges (Waters, Milford, MA) and dried in a SpeedVac (Thermo Electron, Waltham, MA). All chemicals were purchased from SigmaAldrich unless stated otherwise. iTRAQ Labeling

Various amounts of peptides were labeled with 4-plex iTRAQ reagents (Applied Biosystems, Foster City, CA) according to the

TECHNICAL NOTE

instructions from the manufacturer as follows: 200 μg, 114; 100 μg, 115; 200 μg, 116; 100 μg, 117. The labeled peptides were pooled and desalted using Sep-Pak C18 cartridges. They were then fractionated using SCX or ERLIC as follows. SCX Fractionation

The mixture of iTRAQ-labeled peptides was fractionated using a PolySULFOETHYL A column (4.6  200 mm; 5 μm, 200 Å; PolyLC, Columbia, MD) on a Shimadzu Prominence UFLC system. The UV detection was monitored at a wavelength of 214 nm. Buffer A (10 mM KH2PO4 in 25% ACN, pH 3.0) and buffer B (buffer A with 500 mM KCl, pH 3.0) were used to establish a 50 min gradient of 0 5% buffer B for 2 min, 5 20% buffer B for 18 min, 20 40% buffer B for 15 min, and 40 100% buffer B for 10 min, followed by 5 min at 100% buffer B at a flow rate of 1 mL/min with 30 fractions collected. The collected fractions were dried via vacuum centrifuge and combined into 20 fractions as shown in Figure 1A. They were then desalted with Sep-Pak C18 cartridges and redissolved in 3% ACN/0.1% formic acid (FA) for LC MS/MS analysis. ERLIC Fractionation

The mixture of iTRAQ-labeled peptides was fractionated using a PolyWAX LP weak anion-exchange column (4.6  200 mm; 5 μm, 300 Å; PolyLC) on the same UFLC system. The UV detection was monitored at a wavelength of 280 nm. Mobile phase A (10 mM CH3COONH4 in 85% ACN/1% formic acid [FA]) and mobile phase B (30% ACN/0.1% FA) were used to establish a 50 min gradient of 0 28% buffer B for 40 min, 28 100% buffer B for 5 min and 100% buffer B for 5 min at a flow rate of 1 mL/min with 30 fractions collected. The peptides were dissolved in 200 μL mobile phase A for complete solution before being loaded to the column. Mobile phase A was prepared by dissolving CH3COONH4 directly in 85% ACN/1% FA without any adjustment of its pH. The collected fractions were then dried via vacuum centrifuge at 45 °C, combined into 20 fractions as shown in Figure 1B, and redissolved in 3% ACN/0.1% FA for LC MS/MS analysis. LC MS/MS

The peptides were separated and analyzed on a home-packed nanobore C18 column (15 cm  75 μm; 5 μm particles) with a Picofrit nanospray tip (New Objectives, Woburn, MA) on a Tempo nano-MDLC system coupled with a QSTAR Elite Hybrid LC MS/MS system (Applied Biosystems). Peptides from each fraction were analyzed in triplicate by LC MS/MS over a gradient of 90 min. The flow rate of the LC system was set to a constant 300 nL/min. Data acquisition in QSTAR Elite was set to positive ion mode using Analyst QS 2.0 software (Applied Biosystems). MS data was acquired in positive ion mode with a mass range of 300 1600 m/z. Peptides with +2 to +4 charge states were selected for MS/MS. For each MS spectrum, the three most abundant peptides above a five-count threshold were selected for MS/MS and dynamically excluded for 30 s with a mass tolerance of 0.03 Da. Smart information-dependent acquisition was activated with automatic collision energy and automatic MS/MS accumulation. The fragment intensity multiplier was set to 20 and maximum accumulation time was 2 s. Data Analysis

Spectra acquired from the three technical replicates were submitted independently to ProteinPilot (v3.0.0.0, Applied Biosystems) for peak-list generation, protein identification and quantification. User defined parameters of the Paragon algorithm 5569

dx.doi.org/10.1021/pr2007686 |J. Proteome Res. 2011, 10, 5568–5574

Journal of Proteome Research

TECHNICAL NOTE

Figure 1. Chromatograms of (A) SCX and (B) ERLIC fractionation of iTRAQ-labeled peptides. Thirty fractions were collected and then combined into 20 fractions as shown for LC MS/MS analysis. Absorbance was monitored at (A) 214 nm and (B) 280 nm, accounting for the evident difference in sensitivity. The use of 280 nm was necessitated by the absorbency of formic acid in the mobile phase.

in ProteinPilot software were configured as follows: (i) Sample Type, iTRAQ 4-plex (Peptide Labeled); (ii) Cysteine alkylation, MMTS; (iii) Digestion, Trypsin; (iv) Instrument, QSTAR Elite ESI; (v) Special factors, Urea denaturation; (vi) Species, None; (vii) Specify Processing, Quantitate and Bias Correction; (viii) ID Focus, biological modifications, amino acid substitutions; (ix) Database, concatenated IPI rat database (version 3.40; 40381 rsequences) and its reversed complement; (x) Search effort, thorough ID; and (xi) Result quality, Unused ProtScore (Conf) >0.05 (10.0%). Default precursor and MS/MS tolerance for QSTAR ESI MS instrument were adopted automatically by the software. The false discovery rates (FDR) of both peptide and protein identification were set to be less than 1% (FDR = 2.0  decoy_hits/total_hits).21 Theoretical pI values of peptides were calculated on the basis of the algorithm from ENBOSS.22 Peptide GRAVY values were calculated in the same way with the ProtParam tool from SwissProt.23 The average pI and GRAVY values of identified peptides in each fraction were calculated using an in-house Perl program.

’ RESULTS AND DISCUSSION In this study, we extended the benefits of ERLIC fractionation to iTRAQ-labeled peptides in order to enhance proteome coverage and quantitation and to avoid two disadvantages of SCX: the clustering of similarly charged peptides and the need to desalt the fractions. Since SCX-RP is currently the most widely used MDLC for iTRAQ-labeled peptides, we compared it with ERLIC-RP for extent of peptide and protein identification, uniformity of peptide distribution among the fractions, and orthogonality between the two dimensions. The protein extract from rat liver was an appropriate sample for the comparison because of its high complexity. For a fair comparison, SCX and ERLIC fractionation was compared using the same amount of peptides, gradients of identical length, and the collection of an identical number

of fractions. The subsequent LC MS/MS and data analysis methods were also identical. Differences in Mobile Phases Used for the Fractionation of Unlabeled Peptides and iTRAQ-Labeled Peptides

Salt-free mobile phases were employed in the fractionation of unlabeled peptides via ERLIC with good results.20 However, the iTRAQ labeling of the primary amines of peptides greatly reduces their solubility in the same starting solvent (90% ACN/ 0.1% acetic acid). The labeled peptides cannot be completely dissolved even if the concentration of ACN is reduced to 70%. To solubilize the iTRAQ-labeled peptides, 10 mM CH3COONH4 in 85% ACN/1% FA was used as both mobile phase A and the sample solvent for ERLIC fractionation. The addition of a low concentration of volatile ammonium salts and the decrease in ACN concentration enhances the solubility of iTRAQ-labeled peptides. In addition, the volume of mobile phase A used for dissolving peptides before loading should be large enough so that peptides can dissolve completely. The mobile phase B (30% ACN/0.1% FA) was the same as that for unlabeled peptides. Thus, desalting after fractionation was unnecessary. However, vacuum drying at the temperature of 45 °C or higher is necessary to ensure the complete removal of volatile ammonium salts. Combination of Adjacent Fractions Collected in SCX and ERLIC

In a previous study, we found that significantly fewer peptides were identified in fractions at the beginning and end of the SCX gradient,20 and similar results were also reported in another study.13 Thus, as shown in Figure 1A and B, we combined adjacent fractions collected in SCX and ERLIC on the basis of their chromatograms to distribute the peptides evenly into multiple fractions to improve the efficiency of peptide identification and to save instrument time. 5570

dx.doi.org/10.1021/pr2007686 |J. Proteome Res. 2011, 10, 5568–5574

Journal of Proteome Research

TECHNICAL NOTE

Figure 2. Comparison of unique peptide and protein identification from all three SCX-RP and ERLIC-RP replicates.

Table 1. Numbers of Peptide and Protein Identifications (1% FDR) for the SCX-RP and ERLIC-RP Methods peptides

unique peptides

proteins

SCX A

22921

12999

1776

SCX B

21134

12297

1667

SCX C

21571

12382

1804

average of SCX ERLIC A

21875 ( 932 27966

12559 ( 383 16123

1749 ( 72 1968

ERLIC B

32626

18486

2036

ERLIC C

27955

16251

1874

average of ERLIC

29516 ( 2694

16953 ( 1329

1959 ( 81

Protein and Peptide Identifications

LC MS/MS was done in triplicate on peptides in each combined fraction from SCX and ERLIC fractionation. As shown in Figure 2, 2441 proteins and 20 089 unique peptides were identified from the proteome of rat liver from three technical replicates of SCX-RP; 2746 proteins and 30 016 unique peptides were identified from three replicates of ERLIC-RP (FDR < 1%), which were 12.5 and 49.4% higher than those identified in SCX-RP, respectively. In total, 3284 proteins and 38 614 unique peptides were identified from rat liver by a combination of SCXRP and ERLIC-RP. Only 29.8% (11 491/38 614) of the unique peptides were identified by both SCX-RP and ERLIC-RP, indicating that these two methods were highly complementary. The number of peptides, unique peptides, and proteins identified from each of three replicates of SCX-RP and ERLIC-RP are shown in Table 1. Both of them showed excellent reproducibility in identifying proteins and peptides. The detailed list of proteins and unique peptides identified from three replicates of SCX-RP and ERLIC-RP is shown in Supplementary Data 1 4 (Supporting Information). The pI and GRAVY Value Distribution of Peptides

The average number of unique peptides identified in each fraction is illustrated in Figure 3A for SCX-RP and ERLIC-RP. As with the previous report on unlabeled peptides,20 more peptides were identified in most fractions of ERLIC than in SCX. This may reflect better orthogonality between ERLIC and RP. However, significantly more unique peptides were identified in the first and last several fractions of SCX compared with the previous report, indicating the success of the fraction combination strategy shown in Figure 1A. The strategy used for fraction collection can have a pronounced effect on the success of peptide identification. For example, Wang et al. recently reported that a multiple fraction “concatenation” strategy improved peptide and protein

Figure 3. Peptide analysis. (A) Number of unique peptides identified in each fraction of SCX-RP and ERLIC-RP and the average (B) pI and (C) GRAVY value of peptides identified in each fraction of SCX-RP and ERLIC-RP. The labeled standard deviation bars span differences between three replicate runs.

identification greatly in a high-pH RP/low-pH RP sequence.24 To ascertain why ERLIC-RP identifies many more unique peptides and proteins than SCX-RP, the average pI and GRAVY values of all peptides identified in each fraction were calculated and are illustrated for SCX-RP and ERLIC-RP in Figure 3B and C. The average pI of all peptides identified in each fraction is 5.1 7.3 for SCX-RP, a much narrower range than that for ERLIC-RP: 4.0 8.1. This indicates that ERLIC-RP has a better capacity for identification of quite acidic and alkaline peptides than SCX-RP. Poor resolution of peptides with low pI has been reported for SCX fractionation.24 A similar limitation in range of GRAVY values was also observed here for peptides identified by SCX ( 1.13 to 0.17 for SCX-RP vs 1.1 to 1.24 for ERLIC-RP), again consistent with previous reports.20,25 Not only did ERLICRP identify peptides with a wider range of average GRAVY values than did SCX-RP, but it also identified considerably more highly hydrophobic peptides. The average pI and GRAVY value declined gradually with progressive fractions in ERLIC. No trend was obvious for the average GRAVY value in SCX. The disparity between the two modes probably reflects the simultaneous operation of both electrostatic repulsion and hydrophilic interaction in ERLIC, whereas SCX involves electrostatic interaction with no significant role for hydrophilic interaction. The effectiveness of MDLC depends on the chromatographic resolving power of each dimension, as well as on their orthogonality.26 Theoretically, SCX has good orthogonality to RPLC 5571

dx.doi.org/10.1021/pr2007686 |J. Proteome Res. 2011, 10, 5568–5574

Journal of Proteome Research

TECHNICAL NOTE

Table 2. Numbers of Quantified and Altered Proteins from SCX-RP and ERLIC-RP Methods

SCX A SCX B

no. of

quantifications

altered

quantificationsa

with p-value < 0.05

proteinsb

4990 4689

117 121

2 2

SCX C

5052

122

0

in total

14728

360

4

ERLIC A

5438

127

2

ERLIC B

5684

152

0

ERLIC C

5246

128

1

16368

407

3

in total a

If a protein was provided in the ratio of 115:114, 116:114, or 117:114 by Proteinpilot, it was counted as one quantification. Thus, one protein can be quantified three times at most. b Proteins with 1.3-fold changes (>1.3 or