Hydrophilic Interaction Chromatography Using a Meter-Scale

Article pubs.acs.org/ac

Hydrophilic Interaction Chromatography Using a Meter-Scale Monolithic Silica Capillary Column for Proteomics LC-MS Kanta Horie,*,†,‡ Takeo Kamakura,‡ Tohru Ikegami,§ Masaki Wakabayashi,‡ Takashi Kato,† Nobuo Tanaka,§,⊥ and Yasushi Ishihama*,‡ †

Eisai Co., Ltd, Pharmaceutical Science and Technology Core Function Unit, Global Formulation Research, Kawashima, Kakamigahara, Gifu 501-6195, Japan ‡ Graduate School of Pharmaceutical Sciences, Kyoto University, Sakyo-ku, Kyoto, Kyoto 606-8501, Japan § Department of Biomolecular Engineering, Kyoto Institute of Technology, Sakyo-ku, Kyoto, Kyoto 606-8585, Japan ⊥ GL Sciences Inc., 237-2 Sayamagahara, Iruma, Saitama 358-0032, Japan S Supporting Information *

ABSTRACT: A meter-scale monolithic silica capillary column modified with urea-functional groups for hydrophilic interaction liquid chromatography (HILIC) was developed for highly efficient separation of biological compounds. We prepared a ureidopropylsilylated monolithic silica capillary column with a minimum plate height of 12 μm for nucleosides and a permeability of 2.1 × 10−13 m2, which is comparable with the parameters of monolithic silica-C18 capillary columns. Over 300,000 theoretical plates were experimentally obtained in HILIC with a 4 m long column at 8 MPa; this is the best result yet reported for HILIC. A 2 m long ureidopropylsilylated monolithic silica capillary column was utilized to develop a HILIC mode LC-MS system for proteomics applications. Using tryptic peptides from human HeLa cell lysate proteins, we identified the comparable numbers of peptides and proteins in HILIC with those in reversed-phase liquid chromatography (RPLC) using a C18-modified monolithic silica column when shallow gradients were applied. In addition, approximately 5-fold increase in the peak response on average was observed in HILIC for commonly identified tryptic peptides due to the high acetonitrile concentration in the HILIC mobile phase. Since HILIC mode LC-MS shows orthogonal selectivity to RPLC mode LC-MS, it is useful as a complementary tool to increase proteome coverage in proteomics studies.

B

for proteomics, without prefractionation prior to mass spectrometric analysis.10,11 Monolithic silica materials provide high-efficiency separation in a long column format because of their high permeability, and they have been used to achieve high-resolution separations with shallow gradients.6,12−15 Previously, we employed a monolithic silica-C18 capillary column 3.5−4 m in length with an 8−41 h shallow gradient in the RPLC mode for one-dimensional LC separation under non-UHPLC conditions, coupled with an MS/MS system, for analysis of tryptic peptides from 4 μg of

iological mixtures exhibit huge diversity and complexity, and highly efficient separation methods are required to accomplish precise and accurate detection of their components. Recent advances in HPLC, such as the use of ultrahigh pressure liquid chromatography (UHPLC)1−4 or monolithic silica columns,5,6 have considerably increased the separation performance of HPLC methods. The next step for further improvement of the separation power may be the combined use of two different separation modes with distinct physicochemical properties in two-dimensional HPLC (2D-HPLC).7−9 However, a 2D-HPLC system, such as reversed-phase liquid chromatography (RPLC) with prior ion-exchange fractionation, generally needs a long total analysis time and considerable human resources. As an alternative, simple one-dimensional HPLC systems with shallow gradients have also been employed © 2014 American Chemical Society

Received: November 27, 2013 Accepted: March 25, 2014 Published: March 25, 2014 3817

dx.doi.org/10.1021/ac4038625 | Anal. Chem. 2014, 86, 3817−3824

Analytical Chemistry

Article

Escherichia coli proteins and 4 μg of HeLa cell lysate proteins. This approach enabled the identification of 22,196 nonredundant tryptic peptides from 2,602 proteins for Escherichia coli proteins and 41,319 nonredundant tryptic peptides from 5,970 proteins for HeLa cell lysate proteins.16,17 We concluded that the “one-shot proteomics” approach with a long monolithic silica-C18 capillary column and shallow gradient elution in the RPLC mode was very promising as a front-end method for tandem mass spectrometry, but it still remains a challenging proposition to apply it to the more complex human proteome, since the dynamic range for identified human peptides was limited by the presence of coeluted peptides. Thus, it is expected that an improved separation strategy would be effective to overcome this current limitation. To resolve these remaining issues, several researchers have developed approaches that combine hydrophilic interaction liquid chromatography (HILIC) with the orthogonal separation mode of RPLC to reduce the complexity of human proteome analysis.18−21 It was recently reported that the use of particlepacked columns with two different zwitterionic stationary phases in the HILIC mode (ZIC-HILIC and ZIC-cHILIC) prior to RPLC reduced the complexity of peptides in the chromatographic separation window,18 although these approaches were based on offline 2D-HPLC including exhaustive prefractionation, which may reduce quantitative accuracy. Thus, it still is of great interest to evaluate how useful “one-shot proteomics” employing a long monolithic silica capillary column with modified hydrophilic stationary phases in the HILIC mode would be in the proteomics field. In this report, we first describe evaluation of the chromatographic performance of a meter-scale monolithic silica capillary column modified with urea functional groups in the HILIC mode by means of a kinetic approach. Then, we discuss application of the modified column to proteome analysis and further characterization of the column, as well as possible approaches for its optimal usage.

The temperature was then raised slowly (over 10−20 h for long capillary columns), and the monolithic silica columns were incubated for 4 h at 120 °C to form mesopores with the ammonia generated by hydrolysis of urea and then cooled and washed with methanol. After air-drying, the column was heattreated at 330 °C for 24 h to calcinate all organic moieties in the column. Then, these columns were modified with a silylation reagent bearing a urea functional group, UPTMS. The obtained bare-silica monolithic silica capillary column was washed with methanol and then flushed with dried toluene, and a 2:3:3 (v/v/v) or 1:2:2 (w/v/v) mixture of UPTMS, dried pyridine, and dried toluene was passed through the column for 48 h at 80 or 70 °C with nitrogen (1 MPa), respectively, followed by a wash with dried toluene and then methanol for 48 h each via an HPLC pump (10 MPa). This functionbonding step was repeated twice to obtain monolithic silica capillary columns modified with the urea functionality (MS100-UP). After preparation of a long MS-100-UP (approximately 5 m), the column was cut to obtain the desired column length. A commercially available long monolithic silica-C18 capillary column, MonoCap High Resolution 2000 (2 m length ×100 μm I.D.) (MS-100-C18) was donated by GL Sciences (Tokyo, Japan). Evaluation of Fundamental Column Performance. Capillary HPLC measurements to investigate the fundamental chromatographic performance, including van Deemter and kinetics plots, were performed by employing a split flow/ injection HPLC system consisting of an X-LC 3085PU pump (JASCO, Tokyo, Japan), a 7725 injector (Rheodyne, CT) with a splitting T-joint, and a UV detector MU701 with a capillary flow cell connected to an electronic unit via optical fibers (GL Sciences). Conventional HPLC measurements using a particlepacked column were performed with an Agilent 1260 infinity LC system (Agilent Technologies, Santa Clara, CA). Carbohydrates and nucleosides for column characterization were injected directly using a 20 or 50 nL internal-loop sample injector (VICI, Schenkon, Switzerland). During the experiment, the entire system, including the pump, injector, and detector, was kept at 25 °C in a controlled room temperature environment. The UV chromatographic data were collected and processed by EZChrom Elite software (GL Sciences). LC-MS/MS Analyses of HeLa Cell Lysates. HeLa cells (5 × 106) were prepared as described previously24 and digested according to the PTS protocol as described.25 The resulting peptides were desalted with C18-StageTips.26 NanoLC-MS/MS experiments to investigate the chromatographic efficiency including the reproducibility for tryptic peptides were performed on an LTQ (Thermo Fisher Scientific, Bremen, Germany) connected to a Thermo Ultimate3000 nanoflow pump and an HTC-PAL autosampler (CTC Analytics, Zwingen, Switzerland). A TripleTOF 5600 mass spectrometer (AB SCIEX, Foster City, CA) equipped with a Thermo UltiMate 3000 RSLCnano pump and the HTCPAL autosampler was employed for HeLa proteome analysis, as previously described.14,15,17 In both systems, a spray voltage of 2400 V was applied through a PEEK tee connector with a platinum wire. Each coiled 2 m long monolithic silica capillary column was connected to a MonoSpray FS (50 μm I.D. × 50 mm, GL Sciences) or a self-pulled monolithic silica emitter (20 μm I.D. × 50 mm) formed with a Sutter P-2000 (Novato, CA). The flow rate was 500 nL/min. The mobile phases for 2 m long MS-100-UP consisted of (A) 0.5% acetic acid in 10% acetonitrile and (B) 0.5% acetic acid in 95% acetonitrile. A

■

EXPERIMENTAL SECTION Materials. Tetramethoxysilane (TMOS) and methyltrimethoxysilane (MTMS) were obtained from Shin-Etsu Chemicals (Tokyo, Japan), and poly(ethylene glycol) (PEG, MW 10,000) were from Aldrich (St. Louis, MO). Ureidopropyltrimethoxysilane (UPTMS) was obtained from Gelest Inc. (Morrisville, PA). Water purified with a Milli-Q A10 Gradient (Millipore, MA) was used in the experiments. Sequence-grade modified trypsin was obtained from Promega (Madison, WI). All other chemicals and solvents were obtained from Nacalai-Tesque (Kyoto, Japan) and Wako Pure Chemicals (Osaka, Japan) and were used as obtained. Column Preparation. Monolithic silica capillary columns were prepared from a mixture of TMOS and MTMS (v/v 3:1) to form a hybrid structure. The 100 μm I.D. columns were prepared from a mixture of TMOS and MTMS as described in previous reports.22,23 Under typical conditions, the fused-silica capillary tube was first treated with 1 M NaOH at 40 °C for 3 h, followed by a flush with water, and then kept in 1 M HCl at 40 °C for 2 h. After a flush with water and then with acetone, the capillary tube was air-dried at 40 °C. A TMOS/MTMS mixture (9 mL) was added to a solution of PEG (0.9 g) and urea (2.025 g) in 0.01 M acetic acid (20 mL) at 0 °C and stirred for 30 min. The homogeneous solution was then stirred for 10 min at 40 °C, filtered with a 0.45 μm PTFE filter, charged into a fusedsilica capillary tube, and allowed to react at 40 °C overnight. 3818



Article

uηεL (1) P While the chromatographic performance expressed by H shows that the column efficiency of MS-100-UP was similar to that of a 3 μm ZIC-cHILIC column, a much higher permeability was obtained for MS-100-UP (K = 2.1 × 10−13), over 21-fold larger, as compared with columns packed with 3 μm particles (K = 1.0 × 10−14). At a high linear velocity, the columns packed with small particles (e.g., sub-2 μm) could have an advantage in H due to the reduced C term from van Deemter equation; however, high back-pressure is encountered according to eq 1. Apparently, the monolithic silica column has an advantage in permeability compared to the particle-packed columns, as previously reported,16,17,28 even though the surface of the monolithic bare-silica foundation was chemically modified with urea functional groups. Desmet et al. suggested the use of a kinetic plot of t0/N2 against theoretical plate number, N, to evaluate the most important parameters, i.e., maximal achievable theoretical plate number in a certain analysis time and minimal analysis time to achieve a certain theoretical plate number, that enable an HPLC system to operate at optimum performance.30 The kinetic plot is useful to identify the limit of column efficiency under a certain back-pressure. In Figure 1,

two-step linear gradient of 100% to 80% B over 4 or 8 h, 80% to 0% B for 10 min, and 0% B for 10 min was employed throughout this study. One μg of tryptic peptides from HeLa cells was injected into the MS-100-UP. As a control analysis, the mobile phases for 2 m long MS-100-C18 consisted of (A) 0.5% acetic acid and (B) 0.5% acetic acid in 80% acetonitrile. A two-step linear gradient of 5% to 40% B over 4 or 8 h (the same gradient time as MS-100-UP experiments), 40% to 100% B for 5 min, and 100% B for 10 min was employed throughout this study. One μg of tryptic peptides (the same sample volume as MS-100-UP experiments) from HeLa cells was injected into the MS-100-C18. The MS scan range was m/z 300−1500 for all experiments using LTQ and TripleTOF 5600. Regarding the MS conditions of TripleTOF 5600, the MS scans were performed for 0.25 s, and subsequently, 10 MS/MS scans were performed for 0.1 s each. To minimize repeated scanning, previously scanned ions were excluded for 12 s. The collision-induced dissociation (CID) energy was automatically adjusted by the rolling CID function of Analyst TF 1.5. For the investigation of chromatographic performance by using LTQ, the MS conditions were set as previously described.14 Data Analysis for Proteomics Application. The peptides and proteins were identified by Mascot v2.3 (Matrix Science, London, U.K.) against Swiss-Prot database (version 2013_11, 541762 sequences) with a precursor mass tolerance of 20 ppm, a fragment ion mass tolerance of 0.1 Da, and strict trypsin specificity allowing for up to 2 missed cleavages in the experiments using TripleTOF 5600.15,17 Carbamidomethylation of cysteine was set as a fixed modification, and methionine oxidation was allowed as a variable modification.

K=

■

RESULTS AND DISCUSSION Chromatographic Efficiency of Monolithic Silica Capillary Columns Modified with the Urea Functionality. First, the separation characteristics of MS-100-UP were briefly evaluated by using underivatized carbohydrates as representative polar compounds. Bicker et al. have already reported that a particle-packed column modified with the ureidopropyl functional group enables separation in the HILIC mode with an LC-UV system.27 A mixture of acetonitrile and aqueous buffer, frequently employed as a mobile phase in the HILIC mode, is compatible with ESI-MS, so that underivatized carbohydrates can be separated and detected.28 We examined the effect of acetonitrile concentration in the mobile phase on the separation of carbohydrates by MS-100-UP (Supplemental Figure 1, Supporting Information). Retention of the carbohydrates increased drastically with an increase in acetonitrile concentration, indicating that MS-100-UP can be used in the HILIC mode. The chromatographic performance of MS-100-UP was evaluated by a nucleoside, cytidine, and the theoretical plate height, H, was 11−12 μm at the optimum linear velocity in acetonitrile/100 mM ammonium acetate (pH 4.7) (9:1 (v/v)); this is the almost identical efficiency to a well-known commercially available HILIC column, packed with 3 μm particles of ZIC-cHILIC (Merck KGaA, Dermstadt, Germany). Furthermore, the permeability parameter, K, was calculated from the linear velocity (u), viscosity of the mobile phase (η), column porosity (ε), column length (L), and pressure drop (P) as shown below:29

Figure 1. Kinetics plots for various columns under a 20 MPa pressure limit. (a) Solid line: MS-100-UP in acetonitrile/100 mM ammonium acetate (pH 4.7) = 90:10 (v/v) at 25 °C. (b) Dashed line: ZICcHILIC 3 μm column in acetonitrile/100 mM ammonium acetate (pH 4.7) = 80:20 (v/v) at 30 °C. Sample: cytidine (10 μg/mL). Injection volume: 10 μL (1/200 split injection for MS-100-UP). Detection: UV 254 nm.

the kinetic plots for the MS-100-UP and the particle-packed column of 3 μm ZIC-cHILIC are shown for the limiting pressure of 20 MPa, which is appropriate for a conventional HPLC system. As shown in Figure 1, MS-100-UP was found to give higher performance than the particle-packed column in the region of 3 < log(N) at 20 MPa. These results strongly suggest that very high column efficiency in the HILIC mode can be obtained by using a long monolithic column format. In particular, 160,000 and 800,000 theoretical plates were generated with t0 of 1,000 and 10,000 s, respectively, at 20 MPa, i.e., a pressure compatible with a non-UHPLC system. To demonstrate the utility of a long MS-100-UP column, separation chromatograms of nucleobases and nucleosides using MS-100-UP with column lengths of 30 cm and 4 m (100 μm I.D.) are shown in Figure 2A,B, respectively. Excellent 3819



Article

Figure 2. Chromatograms of nucleosides using 30 cm and 4 m MS-100-UP columns. (A) Column: MS-100-UP, 30 cm length × 100 μm I.D. Split flow: ΔP = 0.9 MPa. u = 1.12 mm/s. (B) Column: MS-100-UP, 4 m length × 100 μm I.D. Split flow: ΔP = 7.3 MPa. u = 0.98 mm/s. Mobile phase: acetonitrile/15 mM ammonium acetate (pH 4.7) = 90:10 (v/v). Sample: mixture of ten nucleosides (10 μg/mL each). Injection volume: 50 nL. Detection: UV 254 nm. Peak numbers: 1, toluene; 2, thymine; 3, uracil; 4, 5-methyluridine; 5, adenine; 6, uridine; 7, adenosine; 8, cytosine; 9, guanine; 10, cytidine; 11, guanosine.

long MS-100-UP in the HILIC mode. The separation quality was estimated by examining the elution profile of the tryptic peptides on extracted ion chromatograms (XIC chromatograms), as well as the peak width calculated by the data analysis software. The peaks had full width at half-maximum (fwhm = 2.35 σ) values of approximately 0.4 min as the mean value of major peaks for the identified peptides (SD = 0.3 min, n = 226). During the overall effective separation time period of 4 h (= tg), the separation yielded a peak capacity, PC (= tg/tw, tw: peak width (4σ)), of approximately 360, which was almost equivalent to the previously reported PC value using a long monolithic silica-C18 capillary column in the reversed-phase separation mode.13 We have already reported that a monolithic silica capillary column modified with poly(acrylic acid) in the HILIC mode offered equivalent efficiency to a monolithic silicaC18 capillary column for tryptic peptides from bovine serum albumin, based on PC evaluation.32 In this study, that result was

column efficiencies (over 200,000−300,000 theoretical plates (H = 11−19 μm) for some nucleobases and nucleosides at u = 1 mm/s with the back-pressure of 7 MPa) were observed for 4 m MS-100-UP, resulting in the resolution of compounds that were coeluted from the 30 cm column. These results also indicate that the separation efficiency, expressed by H, is maintained even in the 4 m column format and confirm that homogeneous chemical modification on the long monolithic bed could be achieved. In the case of particle-packed columns for HILIC, it may be difficult to maintain adequate efficiency in the meter-scale capillary format due to the wall effect derived from the slurry packing process,31 as well as the back-pressure problem, as has been widely discussed. Utility of Monolithic Silica Capillary Columns Modified with Urea Functionality for Shotgun Proteomics. Figure 3 shows a base peak chromatogram of tryptic peptides from 1 μg of HeLa cell lysate proteins, obtained from a 2 m 3820



Article

already been established as a front-end method for proteomics.15−17 Figure 4 shows the orthogonality of the two separation modes (HILIC and RPLC) using MS-100-UP and MS-100-

Figure 3. Base peak chromatogram of tryptic peptides from HeLa cell lysate proteins, obtained by HILIC with a 2 m long column of MS100-UP. Column: MS-100-UP, 2 m length × 100 μm I.D. Sample: 1 μg of tryptic peptides from HeLa cell lysate proteins. Gradient time: 4 h. Mass spectrometer: LTQ. For other details, including chromatographic and mass spectrometric conditions, see Experimental Section.

Figure 4. Orthogonality plot of normalized retention times for identified peptides in HILIC with MS-100-UP and RPLC with MS100-C18. n = 4,282 (number of the peptides coidentified in both modes with tg = 8 h, single analysis). Column: MS-100-UP, 2 m length ×100 μm I.D. for HILIC, and MS-100-C18, 2 m length × 100 μm I.D. for RPLC. Sample: 1 μg of tryptic peptides from HeLa cell lysate proteins. Mass spectrometer: TripleTOF 5600. For other details, including chromatographic and mass spectrometric conditions, see Experimental Section.

reproduced even though a different stationary phase bearing a urea functional group was applied to actual human proteome samples. Note that the urea-based column is superior to the poly(acrylic acid)-based columns in terms of the column preparation within the long capillary format, since the clogging problem was not observed during the simple silylation process while it was difficult to prepare the polymer-based columns without clogging owing to the polymerization reaction within the long capillaries. In addition, we repeated the analyses for several months and confirmed the long-term stability of the urea-based column. The retention time reproducibility in the HILIC mode with 4 h gradient was also investigated. For 14 randomly selected tryptic peptides from HeLa proteins, the relative standard deviation (RSD) was less than 2% on average, indicating that the sufficient reproducibility is maintained even though the meter-sacle capillary column was applied to long time analyses in HILIC mode. The number of peptides identified by nanoLC-MS/MS analysis using MS-100-UP was almost equivalent to that using an MS-100-C18 column in the RPLC mode (10,562 peptides/ 2,605 proteins in HILIC, 12,118 peptides/2,529 proteins in RPLC, listed in Supplemental Tables 1−5, Supporting Information) when the same amount of sample (1 μg of tryptic peptides) and gradient times (4 and 8 h) were applied to both columns. Interestingly, a significant number of the identified peptides in the HILIC mode did not overlap with those in the RPLC mode (Supplemental Figure 2, Supporting Information). Approximately 40% of the peptides identified by MS-100-UP in the HILIC mode were unique and distinct from those identified by MS-100-C18 in the RPLC mode. (Approximately 40% of the identified peptides were overlapped within HILIC and RPLC mode.) This result indicates that a focused separation system using MS-100-UP in the HILIC mode would allow identification of additional peptides compared with MS-100-C18 in the RPLC mode, which has

C18 columns. The plots show the normalized retention times (RTi(norm)) of the peptides coidentified in both modes. The normalized retention times (RTi(norm)) were calculated using the following equation: RTi(norm) =

RTi − RTmin RTmax − RTmin

(2)

where RTmax and RTmin were defined as the most and the least retained peptides, respectively.33 As can be seen in the scatter plots, over 50% of separation area was covered by eluting peptides, showing that the retention characteristics of these two modes are orthogonal, as previously reported.33 Application of orthogonal separations should increase the chance of resolving and detecting numerous peptides that would be coeluted in one-mode separation; for instance, RPLC has difficulty in separating peptides with similar hydrophobicity, while HILIC may resolve and detect such peptides based on other characteristics, such as pI as well as hydropathy expressed by the GRAVY (grand average of hydropathy) value.34 In other words, the use of MS-100-UP in the HILIC mode may be a good addition to the use of a meter-scale C18 monolithic silica capillary column in the RPLC mode. Characterization of Monolithic Silica Capillary Columns Modified with Urea Functionality for Proteomics. The GRAVY distributions of peptides eluted in the HILIC and RPLC modes (using MS-100-UP and MS-100-C18, respectively), aligned with the retention times, were evaluated (Supplemental Figure 3, Supporting Information). The GRAVY values of the peptides identified in the HILIC mode were negatively correlated with the retention times, and 3821



Article

workers36,37 were calculated from the retention model (eq 3, where L denotes number of amino acid residues in peptide sequence, bi is the retention coefficient for each type of amino acid residue AAi, and b0 is a linear intercept of the model) and compared to those obtained for MS-100-C18 in RPLC to elucidate the impact of different amino acids on peptide HILIC retention.

hydrophilic peptides tended to be more strongly retained than hydrophobic peptides as expected, whereas a positive correlation was inversely observed in the RPLC mode. Meanwhile, the molecular weight distributions of peptides identified in the HILIC and RPLC modes were almost identical (Supplemental Figure 4, Supporting Information). Figure 5A,B shows the pI distributions of peptides eluted in the HILIC and RPLC modes aligned with the retention times.

RTpred = (1 − 0.21ln L) × (∑ bi × AA i + b0)

(3)

The obtained retention coefficients are shown in Table 1. In contrast to RPLC, the effects of hydrophobic amino acid Table 1. Amino Acid Retention Coefficients for Peptide Separation by HILIC and RPLC retention coefficientsa

Figure 5. Plots of retention times and pI values for identified peptides in the HILIC mode with MS-100-UP and in the RPLC mode with MS-100-C18. n = 4,282 (number of the peptides coidentified in both modes with tg = 8 h, single analysis). Column: MS-100-UP, 2 m length × 100 μm I.D. for HILIC, and MS-100-C18, 2 m length × 100 μm I.D. for RPLC. Sample: 1 μg of tryptic peptides from HeLa cell lysate proteins. Mass spectrometer: TripleTOF 5600. For other details, including chromatographic and mass spectrometric conditions, see Experimental Section.

amino acid residue

MS-100-UP (HILIC)

MS-100-C18 (RPLC)

Ala Arg Asn Asp carbamidomethyl Cys Gln Glu Gly His Ile Leu Lys Met Phe Pro Ser Thr Trp Tyr Val Free term, b0

11.72 −75.62 45.66 162.33 41.42 42.11 101.93 39.44 −55.02 −12.34 −14.11 −85.96 −11.10 −11.16 26.31 45.72 29.91 −8.23 20.95 −2.68 −26.17

41.74 −48.73 12.29 41.60 23.83 21.76 40.80 13.66 −63.12 143.00 159.26 −60.77 120.20 191.49 36.07 20.89 33.91 218.21 99.21 88.37 −156.83

a

Retention coefficients were obtained by regression analysis based on eq 3 for peptides identified from 8 h gradient runs.

residues (Trp, Phe, Leu, and Ile) in HILIC using MS-100-UP were negative or weak for the peptide retention and the presence of hydrophilic and acidic amino acids (e.g., Asp and Glu) residues strongly promoted the peptide retention. Note that hydrophilic and basic amino acids such as His, Lys, and Arg strongly decrease the peptide retention in the HILIC,36,37 supporting that MS-100-UP offers the ERLIC mode separation consisting of electrostatic repulsion as well as hydrophilic interaction for peptides. For the further clarification of charge state at the column surface of MS-100-UP, the method previously reported to characterize the HILIC stationary phase was adopted.38 Xanthine derivatives, theophylline (pKa = 8.6) and theobromine (pKa = 10), were used as test solutes. Theophylline as the more acidic compound was more strongly retained than theobromine, and the obtained chromatogram (data not shown) showed a selectivity factor α (ktheobromine/ ktheophylline) = 0.94, supporting the idea that MS-100-UP is a basic stationary phase. Although the column surface of MS-100UP might be expected to be neutral or slightly negatively charged (acidic) due to the modification with the neutral urea

The pI values of the peptides identified in the HILIC mode were clearly correlated with the retention times, and acidic peptides tended to be more strongly retained than basic peptides, whereas less correlation of pI with retention time was observed in the RPLC mode. These results indicate that the HILIC separation mode using MS-100-UP can be considered as a form of electrostatic repulsion-hydrophilic interaction chromatography (ERLIC), recently introduced by Alpert for the separation of biomolecules and enrichment of phosphopeptides.35 The ERLIC mode is directly applicable to LC-MS/MS, in spite of the electrostatic as well as hydrophilic interactiondriven separation, which is different from the conventional MSincompatible ion-exchange chromatography that is still widely used for 2D-HPLC in the field of proteomics. To probe the characterization of ERLIC for MS-100-UP, the values of amino acid retention coefficients recently proposed by Gilar and co3822



Article

use in the HILIC mode. This approach provided highly orthogonal separation to RPLC with sufficient peak capacity, as well as highly sensitive detection for tryptic peptides. We believe it will be useful as a complementary tool to provide extended proteome coverage in proteomics studies.

functional groups and residual silanol groups, it is interesting that the opposite result was obtained. As discussed above, the developed HILIC mode using MS100-UP is fully compatible with MS and is orthogonal to the existing RPLC, so it appears to have considerable promise for extending the capability of “one-shot proteomics”. It should also be noted that the high organic content of the utilized buffer (0.5% acetic acid in water−acetonitrile at high concentration) can potentially increase the sensitivity of ESI-MS.28,39 Figure 6

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Phone 81-586-89-4728. Fax: 81-586-89-3910. *E-mail: [email protected]. Phone 81-75-7534555. Fax: 81-75-753-4601. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Suguru Ichihara and Ryota Yamana for technical support. We also thank all members of our laboratories for fruitful discussions and GL Sciences for donating a 2 m long monolithic silica-C18 capillary column (MonoCap High Resolution 2000). Thanks are also due to Dr. Tsutomu Setoyama and Dr. Akira Kato in Eisai Co., Ltd, Pharmaceutical Science and Technology Core Function Unit for kind suggestions during this study. This work was supported by JSPS Grants-in-Aid for Scientific Research (No. 24659017, 24116513, and 24241062 to Y.I.).

Figure 6. Comparison of MS intensity for identified peptides in HILIC by MS-100-UP and RPLC by MS-100-C18. n = 4,282 (number of the peptides coidentified in both modes with tg = 8 h, single analysis). Column: MS-100-UP, 2 m length × 100 μm I.D. for HILIC, and MS100-C18, 2 m length × 100 μm I.D. for RPLC. Sample: 1 μg of tryptic peptides from HeLa cell lysate proteins. Mass spectrometer: TripleTOF 5600. For other details, including chromatographic and mass spectrometric conditions, see Experimental Section.

■

REFERENCES

(1) de Villiers, A.; Lauer, H.; Szucs, R.; Goodall, S.; Sandra, P. J. Chromatogr., A 2006, 1113, 84−91. (2) Mellors, J. S.; Jorgenson, J. W. Anal. Chem. 2004, 76, 5441−5450. (3) MacNair, J. E.; Patel, K. D.; Jorgenson, J. W. Anal. Chem. 1999, 71, 700−708. (4) Patel, K. D.; Jerkovich, A. D.; Link, J. C.; Jorgenson, J. W. Anal. Chem. 2004, 76, 5777−5786. (5) Kobayashi, H.; Ikegami, T.; Kimura, H.; Hara, T.; Tokuda, D.; Tanaka, N. Anal. Sci. 2006, 22, 491−501. (6) Luo, Q. Z.; Shen, Y. F.; Hixson, K. K.; Zhao, R.; Yang, F.; Moore, R. J.; Mottaz, H. M.; Smith, R. D. Anal. Chem. 2005, 77, 5028−5035. (7) Horie, K.; Kimura, H.; Ikegami, T.; Iwatsuka, A.; Saad, N.; Fiehn, O.; Tanaka, N. Anal. Chem. 2007, 79, 3764−3770. (8) Shoenmakers, P.; Marriott, P.; Beens, J. LC-GC Eur. 2003, 16, 335−338. (9) Tanaka, N.; Kimura, H.; Tokuda, D.; Hosoya, K.; Ikegami, T.; Ishizuka, N.; Minakuchi, H.; Nakanishi, K.; Shintani, Y.; Furuno, M.; Cabrera, K. Anal. Chem. 2004, 76, 1273−1281. (10) Thakur, S. S.; Geiger, T.; Chatterjee, B.; Bandilla, P.; Frohlich, F.; Cox, J.; Mann, M. Mol. Cell. Proteomics 2011, 10 (8), M110 003699. (11) Zhou, F.; Lu, Y.; Ficarro, S. B.; Webber, J. T.; Marto, J. A. Anal. Chem. 2012, 84, 5133−5139. (12) Eghbali, H.; Sandra, K.; Detobel, F.; Lynen, F.; Nakanishi, K.; Sandra, P.; Desmet, G. J. Chromatogr., A 2011, 1218, 3360−3366. (13) Miyamoto, K.; Hara, T.; Kobayashi, H.; Morisaka, H.; Tokuda, D.; Horie, K.; Koduki, K.; Makino, S.; Nunez, O.; Yang, C.; Kawabe, T.; Ikegami, T.; Takubo, H.; Ishihama, Y.; Tanaka, N. Anal. Chem. 2008, 80, 8741−8750. (14) Horie, K.; Sato, Y.; Kimura, T.; Nakamura, T.; Ishihama, Y.; Oda, Y.; Ikegami, T.; Tanaka, N. J. Chromatogr., A 2012, 1228, 283− 291.

shows a comparison of MS intensity for the peptides identified both in the HILIC mode (MS-100-UP) and RPLC mode (MS100-C18). Surprisingly, despite the same amount of sample loading to MS-100-UP and MS-100-C18, an approximately 5fold increase in the peak response with an average of 4,282 peptides commonly identified was observed in MS-100-UP due to the higher organic content of the mobile phase of HILIC than that in RPLC. Because MS-100-UP in HILIC has smaller sample loading capacity (up to 1 μg) than MS-100-C18 in RPLC (up to 4 μg), the higher sensitivity in HILIC is promising to increase the proteome coverage by HILIC. Further improvement is also expected by optimizing gradient conditions in HILIC since the distribution of identified peptides within the separation window was not equal as shown in Figure 4. It is also noteworthy that relatively basic peptides were selectively eluted in the early phase of the chromatography with high sensitivity and were identified more by this system (Supplemental Figure 5, Supporting Information), since many proteins, especially those found in eukaryotic cell nuclei (such as histones), are highly basic and contain large numbers of post-translational modification sites, which often cause difficulties in shotgun proteomics research.40 These considerations further support the potential value of MS-100UP in the HILIC mode as a complementary strategy in the existing “one-shot proteomics” approach for total proteomics of complex organisms. In conclusion, we have prepared a meter-scale monolithic silica capillary column modified with urea functional groups for 3823



Article

(15) Yamana, R.; Iwasaki, M.; Wakabayashi, M.; Nakagawa, M.; Yamanaka, S.; Ishihama, Y. J. Proteome Res. 2013, 12, 214−221. (16) Iwasaki, M.; Miwa, S.; Ikegami, T.; Tomita, M.; Tanaka, N.; Ishihama, Y. Anal. Chem. 2010, 82, 2616−2620. (17) Iwasaki, M.; Sugiyama, N.; Tanaka, N.; Ishihama, Y. J. Chromatogr., A 2012, 1228, 292−297. (18) Di Palma, S.; Boersema, P. J.; Heck, A. J.; Mohammed, S. Anal. Chem. 2011, 83, 3440−3447. (19) Hao, P.; Guo, T.; Li, X.; Adav, S. S.; Yang, J.; Wei, M.; Sze, S. K. J. Proteome Res. 2010, 9, 3520−3526. (20) McNulty, D. E.; Annan, R. S. Mol. Cell. Proteomics 2008, 7, 971− 980. (21) Boersema, P. J.; Mohammed, S.; Heck, A. J. Anal. Bioanal. Chem. 2008, 391, 151−159. (22) Hara, T.; Kobayashi, H.; Ikegami, T.; Nakanishi, K.; Tanaka, N. Anal. Chem. 2006, 78, 7632−7642. (23) Motokawa, M.; Kobayashi, H.; Ishizuka, N.; Minakuchi, H.; Nakanishi, K.; Jinnai, H.; Hosoya, K.; Ikegami, T.; Tanaka, N. J. Chromatogr., A 2002, 961, 53−63. (24) Sugiyama, N.; Masuda, T.; Shinoda, K.; Nakamura, A.; Tomita, M.; Ishihama, Y. Mol. Cell. Proteomics 2007, 6, 1103−1109. (25) Masuda, T.; Saito, N.; Tomita, M.; Ishihama, Y. Mol. Cell Proteomics 2009, 8, 2770−2777. (26) Rappsilber, J.; Ishihama, Y.; Mann, M. Anal. Chem. 2003, 75, 663−670. (27) Bicker, W.; Wu, J.; Yeman, H.; Albert, K.; Lindner, W. J. Chromatogr., A 2011, 1218, 882−895. (28) Ikegami, T.; Horie, K.; Saad, N.; Hosoya, K.; Fiehn, O.; Tanaka, N. Anal. Bioanal. Chem. 2008, 391, 2533−2542. (29) Neue, U. D. HPLC Columns; Wiley-VCH: New York, 1997. (30) Desmet, G.; Clicq, D.; Gzil, P. Anal. Chem. 2005, 77, 4058− 4070. (31) Bruns, K.; Stoeckel, T.; Smarsly, S.; Höltzel, A.; Tallarek, U. J. Chromatogr., A 2012, 1268, 53−63. (32) Horie, K.; Ikegami, T.; Hosoya, K.; Saad, N.; Fiehn, O.; Tanaka, N. J. Chromatogr., A 2007, 1164, 198−205. (33) Gilar, M.; Olivova, P.; Daly, A. E.; Gebler, J. C. Anal. Chem. 2005, 77, 6426−6434. (34) Kyte, J.; Doolittle, R. F. J. Mol. Biol. 1982, 157, 105−132. (35) Alpert, A. J. Anal. Chem. 2008, 80, 62−76. (36) Gilar, M.; Xie, H.; Jaworski, A. Anal. Chem. 2010, 82, 265−275. (37) Gilar, M.; Jaworski, A. J. Chromatogr., A 2011, 1218, 8890− 8896. (38) Kawachi, Y.; Ikegami, T.; Takubo, H.; Ikegami, Y.; Miyamoto, M.; Tanaka, N. J. Chromatogr., A 2011, 1218, 5903−5919. (39) Naidong, W. J. Chromatogr., B 2003, 796, 209−224. (40) Thomas, C. E.; Kelleher, N. L.; Mizzen, C. A. J. Proteome Res. 2006, 5, 240−247.

3824


Hydrophilic Interaction Chromatography Using a Meter-Scale

Recommend Documents