Preparation of Fe3O4@ZrO2 Core-Shell Microspheres as Affinity Probes for Selective Enrichment and Direct Determination of Phosphopeptides Using Matrix-Assisted Laser Desorption Ionization Mass Spectrometry Yan Li,† Taohua Leng,‡ Huaqing Lin,† Chunhui Deng,*,† Xiuqing Xu,† Ning Yao,† Pengyuan Yang,†,‡ and Xiangmin Zhang*,†
Downloaded via UNIV OF SUNDERLAND on October 9, 2018 at 08:03:58 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Department of Chemistry and Institute of Biomedical Sciences, Fudan University, Shanghai 200433, China Received March 22, 2007
Abstract: Fe3O4@ZrO2 microspheres with well-defined coreshell structure were prepared and applied for the highly selective enrichment of phosphopeptides from tryptic digest product of proteins. To successfully coat iron oxide microspheres with uniform zirconia shell, magnetic Fe3O4 microspheres were first synthesized via a solvothermal reaction, followed by being coated with a thin layer of carbon by polymerization and carbonization of glucose through hydrothermal reaction. Finally, with the use of the Fe3O4@C microspheres as templates, zirconium isopropoxide was prehydrolyzed and absorbed onto the microspheres and eventually converted into zirconia by calcinations. The as-prepared Fe3O4@ZrO2 core-shell microspheres were used as affinity probes to selectively concentrate phosphopeptides from tryptic digest of β-casein, casein, and five protein mixtures to exemplify their selective enrichment ability of phosphopeptides from complex protein samples. In only 0.5 min, phosphopeptides sufficient for characterization by MALDI-MS could be enriched by the Fe3O4@ZrO2 microspheres. The results demonstrate that Fe3O4@ZrO2 microspheres have the excellent selective enrichment capacity for phosphopeptides from complex samples. The performance of the Fe3O4@ZrO2 microspheres was further compared with commercial IMAC beads for the enrichment of peptides originating from tryptic digestion of β-casein and bovine serum albumin (BSA) with a molar ratio of 1:50, and the results proved a stronger selective ability of Fe3O4@ZrO2 microspheres over IMAC beads. Finally, the Fe3O4@ZrO2 microspheres were successfully utilized for enrichment of phosphopeptides from human blood serum without any other purification procedures. Keywords: Fe3O4@ZrO2 microspheres • Selective enrichment • MALDI-MS • Phosphopeptides
* To whom correspondence should be addressed. E-mails: (C.D.)
[email protected]; (X.Z.)
[email protected]. Tel: +86-21-65643983. Fax: +86-21-6564-1740. † Department of Chemistry, Fudan University. ‡ Institute of Biomedical Sciences, Fudan University.
4498
Journal of Proteome Research 2007, 6, 4498-4510
Published on Web 09/27/2007
Introduction Phosphorylation is one of the most important post-translational protein modifications in living cells, and its investigation is of key interest in the field of proteomics. Generally, the enzymatic digest products of phosphoproteins can be characterized using matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS). However, the signals of the phosphopeptides are often suppressed by nonphosphorylated peptide residues contained in protein digest. Therefore, selective enrichment of phosphopeptides from enzymatic digest product of proteins remains an interesting research topic in the filed of bioanalytical chemistry. A number of strategies have been applied to separate phosphorylated proteins and peptides from the nonphosphorylated ones. One such strategy involves the antibodies specific for phosphorylated amino acids1-3 which is limited due to the availability, high costs, and specificity of the antibodies. Another commonly used strategy is immobilized metal ion affinity chromatography (IMAC) which relies on the affinity of the phosphate group to metal ions (Fe3+ or Ga3+) immobilized on a matrix such as agarose via acidic compounds like iminodiacetic acid (IDA) or nitrilotriacetic acid (NTA).4-9 Acidic phosphopeptides retained on porous resin beads can be directly characterized by MALDI-TOF-MS, which allows sequencing of the enriched phophopeptides without the necessity for prior elution of the peptides from the bead, thus, avoiding sample loss during the elution process. However, due to the ‘shadow effect’10 of porous beads, a large number of phosphopeptides bound in the pores are not fully accessible during direct laser desorption. During the past decades, metal oxide affinity chromatography (MOAC) has gained increased attention. Metal oxides such as titania-11-13 or aluminacoated14-17 adsorbents have been demonstrated to be effective materials for the enrichment of phosphopeptides as they seem to have higher selectivity for phosphopeptides because of their reduced nonspecific binding compared to IMAC and the results are more reproducible. Recently, Håkansson’s group18 has demonstrated the enrichment of phosphorylated peptides on the surface of zirconium dioxide loaded in a microtip. According to their statements, zirconium oxide microtips have similar performance to titanium oxide microtips but possess a unique selectivity for singly phosphorylated peptides. Magnetic materials, such as superparamagnetic iron oxide particles, have been extensively used for both in vitro and in 10.1021/pr070167s CCC: $37.00
2007 American Chemical Society
technical notes vivo application, such as cell separation,19,20 magnetically assisted drug delivery,21,22 and enzyme immobilization.23,24 Additionally, magnetic materials are the most commonly selected substrates as affinity probes because of the ease of isolation of the magnetic-material-target conjugate from the sample solution based on their magnetic properties.9 Affinity probes with magnetic properties lead the isolation of the trapped species from sample solutions very conveniently and rapidly; therefore, the time required for separation can be dramatically reduced. Recently, Chen et al.25 synthesized ZrO2coated magnetic particles (not core-shell structure), and used them as the rapid trapping adsorbents for the analysis of phosphopeptides. However, the ZrO2-coated magnetic particles they synthesized had the disadvantage of nonspecific binding of nonphosphopeptides.25 We speculated that Fe3O4@ZrO2 microspheres with core-shell structure may overcome this shortcoming. In this study, we present a novel method to prepare Fe3O4@ZrO2 microspheres with well-defined core-shell structure and apply them as affinity probes for the highly selective enrichment of phosphopeptides from tryptic digest product of proteins.
Experimental Section 1. Preparation and Characterization of Fe3O4@ZrO2 CoreShell Microspheres. 1.1. Synthesis of Fe3O4@C Magnetic Microspheres. Fe3O4 microspheres with diameter of 280 nm were first synthesized via a solvothermal reaction as previously described.9 In the next step, 0.05 g of microspheres were ultrasonicated for 10 min in 0.1 M HNO3, followed by washing with deionized water. Then, the treated Fe3O4 microspheres were redispersed in 0.5 M aqueous glucose solution. After vigorous stirring for 10 min, the suspension was transferred to autoclaves and kept at 180 °C for 4 h. After reaction, the autoclaves were cooled naturally in air, and the suspensions were isolated with the help of a magnet and washed with deionized water and alcohol three times, respectively. The final sample was obtained after oven-drying at 80 °C for more than 4 h. 1.2. Preparation of Fe3O4@ZrO2 Core-Shell Microspheres. Zirconium isopropoxide (0.3 g) was dissolved in ethanol (50 mL) to form a clear solution. Fe3O4@C magnetic microspheres (∼100 mg) were then dispersed in the freshly prepared solution with the aid of ultrasonication for 5 min. A 1:5 (v/v) mixture of water and ethanol was added dropwise to the suspension of Fe3O4@C magnetic microspheres with vigorous magnetic stirring over a period of approximately 15 min. Thereafter, the suspension was stirred for a further 1 h before separation and washing with ethanol. After five cycles of separation/washing/ redispersion with ethanol, the powder obtained was oven-dried and calcined in air at 500 °C for 1 h. The furnace was then left to cool to room temperature. 1.3. Characterization. The transmission electron microscope (TEM)characterizedFe3O4 microspheres,Fe3O4@C,andFe3O4@ZrO2 core-shell microspheres. The minimum size of the electron beam was ca. 0.7 nm. The Fourier transform infrared (FT-IR) spectra of Fe3O4 microspheres, Fe3O4@C, and Fe3O4@ZrO2 core-shell microspheres were recorded on FT-IR (NEXUS470, NICOLET). 2. Protein Enrichment Experiment. 2.1. Materials and Sample Preparation. Bovine β-casein and casein (from bovine milk), chicken egg albumin (ovalbumin), myoglobin horse heart, cytochrome C, bovine serum albumin (BSA), trypsin
Li et al.
(from bovine pancreas, TPCK treated), phosphoric acid (H3PO4), ammonium bicarbonate (NH4HCO3), and 2,5-dihydroxybenzoic acid (2, 5-DHB) were purchased from Sigma Chemical (St. Louis, MO). Acetonitrile (ACN) was purchased from Merck (Darmstadt, Germany). All aqueous solutions were prepared using Milli-Q water by Milli-Q system (Millipore, Bedford, MA).All other chemicals and reagents were of the highest grade commercially available. For bovine β-casein, casein (from bovine milk), chicken egg albumin (ovalbumin), myoglobin horse heart, cytochrome C, bovine serum albumin (BSA), each protein was dissolved in 25 mM ammonium bicarbonate buffer at pH 8.0 with a final concentration of 1 mg/mL and treated with trypsin (2%, w/w) for 12 h at 37 °C. 2.1.1. Peptide Mixture 1. Peptide mixture 1 contained peptides originating from a tryptic digestion of 2 × 10-8 M of β-casein. 2.1.2. Peptide Mixture 2. Peptide mixture 2 contained peptides originating from a tryptic digestion of 5ng/µL of commercial casein. 2.1.3. Peptide Mixture 3. Peptide mixture 3 contained peptides originating from tryptic digestions of 2 × 10-8 M of β-casein, ovalbumin, myoglobin horse heart, cytochrome C, and bovine serum albumin (BSA). 2.1.4. Peptide Mixture 4. Peptide mixture 4 contained peptides originating from tryptic digestion of 50 nM of β-casein and 2.5 µM of BSA which results in a molar ratio of 1:50. All the Peptide mixtures were diluted with 50% acetonitrile and 0.1% TFA aqueous solution (v/v). Blood samples obtained from Zhongshan Hospital were collected in 8.5 mL, allowed to clot at room temperature for up to 1 h, and centrifuged at 4 °C for 5 min at 1000 rpm. Sera (upper phase) were aliquoted and stored frozen at -80 °C. Before use, 10 µL of serum sample was diluted with 20 µL of 50% acetonitrile (0.15% TFA) aqueous solution (v/v), and without any other purification and tedious treatment, the serum sample was ready for enrichment. 2.2. Enrichment of Phosphopeptides by Fe3O4@ZrO2 CoreShell Microspheres. A suspension of Fe3O4@ZrO2 core-shell microspheres (5 µL of 20 mg mL-1) was added into 200 µL of peptide mixture originating from tryptic digestions or 30 µL of diluted serum sample, respectively. Then the mixed solutions were vibrated at 37 °C for several minutes. After that, with the help of magnet, the peptides-loaded Fe3O4@ZrO2 core-shell microspheres were collected by removal of the supernatant and washed with 50% acetonitrile and 0.1% TFA aqueous solution (v/v) three times. Then, the obtained peptides-loaded Fe3O4@ZrO2 microspheres were redispersed in 10 µL of 50% acetonitrile aqueous solution (v/v). 2.3. Enrichment of Phosphopeptides by IMAC Beads. IMAC purification of phosphorylated peptides was performed according to Larsen et al.26 with minor changes. Briefly, 80 µL of iron-coated PHOS-select metal chelate beads (Sigma) was washed two times in 400 µL of washing/loading solution (0.25 M acetic acid and 30% acetonitrile), resuspended in 200 µL of peptide solution, and incubated for 30 min with constant rotating. After incubation, the solution was loaded onto a constricted GeLoader tip, and a gentle air pressure was used to pack the beads. Subsequently, the beads were washed extensively with the washing/loading solution. The bound peptides were eluted using 10 µL of NH4OH (pH 10.5) and neutralized with formic acid prior to MALDI-MS analysis. Journal of Proteome Research • Vol. 6, No. 11, 2007 4499
Fe3O4@ZrO2 Core-Shell Microspheres as Affinity Probes
technical notes
Figure 1. TEM images of (a) Fe3O4 microspheres, (b) Fe3O4@C microspheres, (c) Fe3O4@ZrO2 core-shell microspheres. (d) The EDX spectrum data of the obtained Fe3O4@ZrO2 core-shell microspheres.
2.4. MALDI-TOF-MS Process. The above-mentioned, peptides-loaded Fe3O4@ZrO2 microspheres’ slurry was deposited on the MALDI target using dried droplet method. A total of 0.5 µL of the slurry was deposited on the plate and then another 0.5 µL of a mixture of 20 mg mL-1 2,5-dihydroxybenzoic acid (in 50% acetonitrile aqueous solution, v/v) and 1% (v/v) H3PO4 aqueous solution, 1:1 (v/v), was introduced as a matrix. MALDI-TOF-MS experiments were performed in positive ion mode on a 4700 Proteomics Analyzer (Applied Biosystems) with the Nd:YAG laser at 355 nm, a repetition rate of 200 Hz, and an acceleration voltage of 20 kV.
Results and Disucussion To coat uniform zirconia on single magnetite microsphere, the as-made Fe3O4 microspheres (Figure 1a) were first coated with hydrophilic carbon by hydrothermal reaction of glucose, which results in core-shell-structured Fe3O4@C microspheres (Figure 1b) with uniform carbon layer of about 20 nm in thickness. The TEM image of Fe3O4@ZrO2 microspheres obtained by calcination of Fe3O4@C microspheres with absorbed zirconia oligomers (Figure 1c) shows that the diameter of the microspheres is ∼280 nm, and the typical core-shell structure can be clearly observed, indicating the successful coating of ZrO2 on Fe3O4. Notably, as can be seen in the TEM image, the outer shell of the obtained Fe3O4@ZrO2 core-shell micro4500
Journal of Proteome Research • Vol. 6, No. 11, 2007
spheres (∼20 nm) consists of larger amount of zirconia nanoparticles, which may provide a highly specific area and is beneficial for specifically trapping phosphopeptides when they are used for enrichment application. The energy-dispersive X-ray analysis (EDXA) (Figure 1d) of the obtained Fe3O4@ZrO2 by illuminating electron beams on the obtained core-shell microspheres reveals the existence of Fe, Zr, and O elements, further confirming the formation of zirconia on the Fe3O4 microspheres. To further confirm that we had successfully prepared Fe3O4@ZrO2 core-shell magnetic microspheres, we employed FT-IR spectroscopy to examine the surfaces of the as-made Fe3O4, Fe3O4@C, and Fe3O4@ZrO2 microspheres. The characteristic band of Fe3O49 appears at ∼576 cm-1 (Figure 2a). After a coating with a thin layer of C, the Fe3O4@C microspheres show bands at 1700 and 1625 cm-1 associated with the CdO vibration and CdC vibration, respectively, indicating the carbonization of glucose during hydrothermal reaction (Figure 2b). The peaks at 1232, 1292, and 1352 cm-1 are attributed to the C-O stretching and O-H bending vibrations, suggesting the presence of large amount of hydrophilic groups. The presence of these hydrophilic groups not only endows Fe3O4@C microspheres with better dispersibility and stability than those of the as-synthesized Fe3O4 microspheres, but also significantly enhances the affinity between the microspheres and the
technical notes
Li et al.
Figure 2. FI-TR spectra of (a) as-synthesized Fe3O4 microspheres, (b) Fe3O4@C microspheres, and (c) Fe3O4@ZrO2 core-shell microspheres.
prehydrolyzed zirconium isopropoxide (i.e., zirconia oligomers). The FT-IR spectrum of the Fe3O4@ZrO2 microspheres (Figure 2c) shows characteristic bands at 634 cm-1 of zirconia. These FT-IR spectroscopy results, therefore, provide additional evidence for the successful synthesis of Fe3O4@ZrO2 core-shell microspheres.
To investigate the selectivity and efficiency of the obtained Fe3O4@ZrO2 microspheres for enrichment of phosphopeptides, the bovine β-casein was first chosen as a model phosphoprotein due to its well-characterized phosphorylation sites.27 Table 1 lists the sequences of the phosphopeptide ions derived from β-casein. Two are monophosphorylated peptides, with theoretical m/z (mass/charge) values of 2061.83 and 2556.09, respectively. As they contain only one phosphoserine, they can be directly detected by MALDI-TOF-MS. Another is the tetraphosphorylated peptide, RELEELNVPGEIVE[pS]L[pS][pS][pS]EESITR, with a theoretical m/z value of 3122.27, which is large and difficult to ionize. Figure 3a presents the direct MALDI mass spectrum of the tryptic digest of β-casein (2 × 10-7 M) prior to enrichment. Although the peaks at m/z 1137.54, 1195.65, 1252.61, 1368.67, 1383.78, 2061.76, 2186.12, 2223.14, and 3122.31 are present in this mass spectrum, only the two weak peaks with m/z at 2061.76 and 3122.31 represent phosphopeptide residues derived from β-casein. Acidified tryptic β-casein digest (at certain pH) was mixed with the Fe3O4@ZrO2 core-shell microspheres and was incubated for several minutes; then, simply with the help of magnet, the phosphopeptides-loaded magnetic microspheres were isolated, washed, and then prepared for MS analysis using 1% (v/v) phosphoric acid and 2,5-dihydroxybenzoic acid (in 50% acetonitrile aqueous solution, v/v), 1:1 (v/v), as a matrix. Zirconium oxide is known to have amphoteric properties; that is, it can react either as a Lewis acid or base depending on the pH of the reaction solution. In acidic solution, ZrO2 behaves as a Lewis acid with positively charged zirconium atoms, thereby displaying anion-exchange properties.18,28 The binding constant of phosphate ions is markedly higher than that for other Lewis bases,29,30 suggesting that high binding selectivity for phosphorylated peptides over nonphosphorylated acidic peptides should be achievable with ZrO2 upon careful selection of the pH. As demonstrated in Figure 3b and Table 1, after enrichment using the Fe3O4@ZrO2 core-shell microspheres as affinity probes for 0.5 min at pH 2, not only three peaks belonging to phosphopeptides of β-casein, marked 2061.97, 2556.21, and 3122.38, appeared with much higher intensity, but also the dephosphorylated fragments of phosphopeptides were efficiently enriched. The signal at m/z 1978.99 could be assigned to a dephosphorylated fragment of phosphopeptide through loss of phosphoric acid, due to the metastable loss of phosphoric acid from the parent ions; the difference between 2061.97 and 1978.99 is 82.98 Da instead of 98 Da (1 Da ) 1.66 × 10-27 kg).31 The peak marked with 1031.57 could be assigned to double-charged ion of monophosphorylated peptide derived from β-casein. In addition, since the β-casein sample we used is only 90% pure, it contains some R-casein. The peaks at m/z 1466.76, 1660.98, and 1952.08 represent phosphopeptide residues derived from R-casein (see Table 1). Compared to the tryptic β-casein digest prior to enrichment (Figure 3a), almost
Table 1. Phosphopeptides Identified in Tryptic Digest of β-Casein after Enrichment by Fe3O4@ZrO2 Core-Shell Microspheres
a
aa
peptide sequences
theoretic m/z
observed m/z
β/33-48 β/33-52 β/1-25 R-S2/138-149 R-S1/106-119 R-S1/104-119
FQ[pS]EEQQQTEDELQDK FQ[pS]EEQQQTEDELQDKIHPF RELEELNVPGEIVE[pS]L[pS][pS][pS]EESITR TVDME[pS]TEVFTK VPQLEIVPN[pS]AEER YKVPQLEIVPN[pS]AEER
2061.83 2556.09 3122.27 1466.51 1660.79 1951.95
2061.97 2556.21 3122.38 1466.76a 1660.98a 1952.08a
Phosphopeptides derived from R-casein.
Journal of Proteome Research • Vol. 6, No. 11, 2007 4501
Fe3O4@ZrO2 Core-Shell Microspheres as Affinity Probes
technical notes
Figure 3. (a) Direct MALDI mass spectrum of the tryptic digest product of β-casein (2 × 10-7 M) without any enrichment. MALDI mass spectra of on-bead analysis of phosphopeptides obtained when using Fe3O4@ZrO2 core-shell microspheres to selectively trap target peptides from the tryptic digest product of β-casein (2 × 10-7 M) for 0.5 min at various binding pH values: (b) pH 2, (c) pH 4, (d) pH 6. Phosphopeptide ions derived from β-casein and R-casein are marked with asterisks and pound sign, respectively. The metastable losses of phosphoric acid are indicated with 4. The data in parentheses are S/N of the corresponding peaks. 4502
Journal of Proteome Research • Vol. 6, No. 11, 2007
technical notes
Li et al.
Figure 4. MALDI mass spectra of on bead analysis of phosphopeptides obtained when using Fe3O4@ZrO2 core-shell microspheres to selectively trap target peptides from the tryptic digest product of β-casein (2 × 10-7 M) at pH 2 with incubation time of (a) 5 min, (b) 15 min, and (c) 60 min. Phosphopeptide ions derived from β-casein and R-casein are marked with asterisks and pound sign, respectively. The metastable losses of phosphoric acid are indicated with 4. The data in parentheses are S/N of the corresponding peaks.
no unphosporylated peptides were seen in Figure 3b, and the signal-to-noise ratio (S/N) of the MS signal of the phosphorylated peptides enriched by Fe3O4@ZrO2 core-shell microspheres was significantly improved. When binding at pH 4 (Figure 3c), phosphopeptides still dominate the mass spectrum; however, some nonphosphorylated peptides such as 1137.77, 1252.82, 1761.11, and 2163.21 were observed in the spectrum, indicating the compromising selectivity at higher pH. At pH 6 (Figure 3d) at which the Lewis acid property of ZrO2 diminishes more, phosphopeptides are no longer the dominant species and a limited amount of phosphopeptides binds to the Fe3O4@ZrO2 core-shell microspheres. These data show that it
is critical to use low pH solutions to achieve high phosphopeptide binding selectivity. We further investigated whether the results were improved by extending the enrichment time to 5, 15, and 60 min. Figure 4 displays the MALDI mass spectra obtained after incubating the tryptic digest product of β-casein with Fe3O4@ZrO2 coreshell microspheres under vortex mixing for 5, 15, and 60 min. The same three peaks derived from phosphopeptides of β-casein appeared in the mass spectra, but when incubation time increases, signal abundance of peaks 2556.09 and 3122.27 becomes higher. Nevertheless, the results still indicated that phosphopeptides sufficient for MALDI-MS analysis could be Journal of Proteome Research • Vol. 6, No. 11, 2007 4503
Fe3O4@ZrO2 Core-Shell Microspheres as Affinity Probes
technical notes
Figure 5. MALDI mass spectra of (a and c) the tryptic digest product of β-casein without enrichment: (a) 2 × 10-8 M, 0.5 µL; (c) 2 × 10-9 M, 0.5 µL. (b and d) On bead analysis of phosphopeptides obtained when using Fe3O4@ZrO2 core-shell microspheres to selectively trap target peptides from the tryptic digest product of β-casein at the following concentrations and extraction volumes: (b) 2 × 10-8 M, 200 µL; (d) 2 × 10-9M, 200 µL. The data in parentheses are S/N of the corresponding peaks. 4504
Journal of Proteome Research • Vol. 6, No. 11, 2007
technical notes enriched within only 0.5 min. Even when the concentration of the sample is lowered to 2 × 10-8 and 2 × 10-9 M, as shown in Figure 5b,d, phosphopeptides at theoretical m/z’s of 2061.83 and 2556.09 were still observed in the mass spectra after enrichment. The reason the tetraphosphorylated peptide at m/z 3122.27 was not observed in the mass spectra (Figure 5b,d) may be due to two factors, that is, ionization efficiency and selective binding affinity. Below certain concentrations, the ions of multiply phosphorylated peptides are easily suppressed by the presence of singly phosphorylated peptides because of to their poorer ionization efficiency. Additionally, as demonstrated in Håkansson’s work,18 ZrO2 possesses a unique selectivity for singly phosphorylated peptides. Therefore, when samples are at low concentration, the peaks derived from single-phosphorylated peptides dominate the mass spectra after enrichment by the Fe3O4@ZrO2 core-shell microspheres. The results above indicate that the selectivity and affinity of the Fe3O4@ZrO2 core-shell microspheres for phosphopeptides are quite good. Furthermore, within only 0.5 min, phosphopeptides in quantities sufficient for MALDI-MS analysis are enriched, indicating excellent enrichment capacity of Fe3O4@ZrO2 core-shell microspheres even at a very low level. Quantification of the binding and recovery of phosphopeptides to the Fe3O4@ZrO2 core-shell microspheres was investigated and compared to the commercial IMAC beads. Peptides originating from tryptic digest of β-casein (4 pmol) were enriched using Fe3O4@ZrO2 core-shell microspheres and commercial IMAC resin, respectively. After enrichment, the phosphopeptide-conjugated microspheres were redispersed in 10 µL of loading buffer, and 0.5 µL of the dispersion was directly deposited onto the MALDI plate for MS analysis. In a parallel experiment, the phosphopeptides conjugated to the microspheres were eluted with 10 µL of NH4OH (pH 10.5) for 5 min. The elution was then neutralized with formic acid, and 0.5 µL of them was used for MALDI-MS analysis. When using IMAC resin, the phosphopeptides were also eluted with NH4OH and neutralized prior to MALDI-MS analysis (as described in Experimental Section). The peak intensities of two monophosphopeptides and one tetraphosphopeptide with molecular weights of 2061.8, 2556.2, and 3122.1, respectively, were used in the assay for comparison. The data in Figure 6 show that either monophosphorylated peptides or multiphosphorylated peptides can be directly analyzed using the magnetic microspheres without elution, even though the ion signals were suppressed to a certain extent. This may be due to the introduction of solid particles on the MALDI plate causing inhomogeneity of sample cocrystallization with the matrix. As shown in Figure 6, phosphopeptides were efficiently eluted from the Fe3O4@ZrO2 core-shell microspheres, except that the peak intensity of tetraphosphopeptide is lower than that with direct analysis and IMAC, indicating a relatively poorer elution of multiphosphorylated peptides from Fe3O4@ZrO2 core-shell microspheres with NH4OH (pH 10.5). However, we believe that, with optimized elution parameters, efficient recovery of the multiphosphorylated peptides can be enhanced. In contrast, the IMAC beads demonstrated better recovery for the multiphosphorylated peptides but worse recovery for the monophosphopeptides compared to the Fe3O4@ZrO2 core-shell microspheres. Selectivity of these magnetic microspheres for phosphopeptides was further demonstrated with a tryptic digest of the mixture of casein (composed of R-S1 and R-S2 units and β-casein), at a low concentration (5 ng/ µL). Figure 7a presents
Li et al.
Figure 6. Quantification of the binding and recovery of phosphopeptides to the Fe3O4@ZrO2 core-shell microspheres and commercial IMAC beads. Peptides originating from tryptic digest of β-casein (4 pmol) were used for enrichment. ‘Direct analysis’ means an aliquot of suspension of phosphopeptide-conjugated microspheres were directly deposited onto the MALDI plate for MS analysis. The phosphopeptides conjugated to the Fe3O4@ZrO2 core-shell microspheres and IMAC bead were eluted with 10 µL of NH4OH (pH 10.5) for 5 min, respectively. The peak intensities of two monophosphopeptides and one tetraphosphopeptide with molecular weights of 2061.8, 2556.2, and 3122.1, respectively, were used in the assay for comparison. Further experiment details are to be found in Experimental Section.
the direct MALDI mass spectrum of the tryptic digest of casein (5 ng/ µL) prior to enrichment. Among the peaks, only four phosphorylated peptide ions with low intensity were observed, marked with asterisks at m/z no. 2 1466.67 (S2/138-149), no. 4 1660.86 (S1/106-119), and no. 8 2061.87 (β/33-48). After the affinity probes are used to enrich phosphopeptides, phosphopeptide residue ions start to appear in the MALDI mass spectrum (Figure 7b and Table 2). The peaks marked with the nos. 4, 5, 6, 7, and 11 are derived from R-S1-casein, whereas the peaks marked with the nos. 1, 2, 3, and 10 are derived from R-S2-casein. The remaining peaks, marked with the nos. 8, 9, and 12, are derived from β-casein. The corresponding peptide sequences of these ions are listed in Table 2. The peaks at m/z 1867.09 and 1977.83 could be assigned to dephosphorylated fragments of phosphopeptide through the metastable loss of phosphoric aicd from peak 7 (m/z 1951.71) and peak 8 (m/z 2061.56), respectively. Nearly all of the ions revealed in the mass spectrum are phosphopeptide residues. Although there are more than 10 phosphorylated peptides in the same sample solution, there is no serious suppression effect between them when trapping and analyzing, meaning that all the phosphopeptide residues can be trapped simultaneously with no interference. The results indicate that the Fe3O4@ZrO2 coreshell microspheres can be used to selectively enrich phosphopeptides from a complex sample. This shows that Fe3O4@ZrO2 core-shell microspheres have better selectivity for phosphopeptides than the ZrO2-coated Fe3O4 microspheres.25 To evaluate the ability of the Fe3O4@ZrO2 core-shell microspheres to capture the phosphopeptides from an even more complicated sample, we mixed the tryptic digest products of five proteins, including nonphosphorylated proteins (cytochrome C, myoglobin, and bovine serum albumin (BSA)) and Journal of Proteome Research • Vol. 6, No. 11, 2007 4505
Fe3O4@ZrO2 Core-Shell Microspheres as Affinity Probes
technical notes
Figure 7. MALDI mass spectra of (a) 5 ng/ µL tryptic casein digest without any enrichment; (b) on-bead analysis of phosphopeptides obtained when using Fe3O4@ZrO2 core-shell microspheres to selectively trap target peptides from the 5 ng/ µL tryptic digest product of casein, 200 µL. The phosphopeptides are marked with numbers. Table 2. Phosphopeptides Identified in Tryptic Digest of Casein after Enrichment by Fe3O4@ZrO2 Core-Shell Microspheres no.
aa
peptide sequences
observed m/z
1 2 3 4 5 6 7 8 9 10 11 12
R-S2/138-147 R-S2/138-149 R-S2/126-137 R-S1/106-119 R-S1/104-119 R-S1/43-58 R-S1/104-119 β/33-48 β/33-52 R-S2/2-21 R-S1/99-120 β/1-25
TVDME[pS]TEVF TVDME[pS]TEVFTK EQL[pS]T[pS]EENSKK VPQLEIVPN[pS]AEER YKVPQLEIVPN[pS]AEER DIG[pS]E[pS]TEDQAMETIK YKVPQLEIVPN[pS]AEER FQ[pS]EEQQQTEDELQDK FQ[pS]EEQQQTEDELQDKIHPF NTMEHV[pS] [pS] [pS]EESII[pS]QETYK LRLKKYKVPQLEIVPN[pS]AEERL RELEELNVPGEIVE[pS]L[pS][pS][pS]EESITR
1237.36 1466.47 1561.97a 1660.61 1832.62 1927.46 1951.71 2061.56 2555.74 2618.62 2703.56 3121.94
a
Phosphopeptides ion peak is [M + Na]+.
phosphoproteins (β-casein and chicken egg albumin (ovalbumin)). Figure 8a displays the direct MALDI mass spectrum of the tryptic digest product (0.5 µL) of cytochrome C, myoglobin, BSA, β-casein, and ovalbumin, which are all of 2 × 10-8 M. At such a low concentration, only two weak phosphopeptide residue ions are observed in the MALDI mass spectrum prior to enrichment. The remaining peaks are all nonphosphorylated peptide residues derived from these proteins. However, after enrichment, 8 phosphopeptide residue ion peaks appear in the mass spectrum (as shown in Figure 8b). Among these peaks, m/z at 2061.92, 2556.21, and 3122.36 are derived from β-casein, and the peak at m/z 1967.08 could be assigned to a dephosphorylated fragment of phosphopeptide through the metastable loss of phosphoric acid from the parent ions. The peak marked 4506
Journal of Proteome Research • Vol. 6, No. 11, 2007
with 1031.57 could be assigned to double-charged ion of monophosphorylated peptide derived from β-casein. The phosphopeptide marked with 2088.99 is derived from ovalbumin, and with the metastable loss of phosphoric acid, a peak at m/z 1994.29 appears in the mass spectrum. Because of the impurity of the β-casein we used in the experiment, a peak with m/z of 1466.68 could also be observed, which is a monophosphorylated peptide from R-casein, TVDME[pS]TEVFTK. Furthermore, there are no nonphosphorylated peptides appearing in the mass spectrum (Figure 8b). The results further demonstrate that, compared with the simple ZrO2coated Fe3O4,25 the Fe3O4@ZrO2 core-shell microspheres are very effective in selectively trapping phosphopeptides from a very complex sample. We have also compared the performance of our Fe3O4@ZrO2 core-shell microspheres with commercially available IMAC materials (PHOS-select iron affinity beads, Sigma) for selective enrichment from a semicomplex sample (tryptic-digested peptides of β-casein and bovine serum albumin (BSA) with a molar ratio of 1:50). A direct analysis of tryptic digestion of 0.5 µL of Peptide mixture 1 by MALDI-MS was carried out, and the obtained mass spectrum was shown in Figure 9a. The phosphopeptides from β-casein were difficult to distinguish from the mass spectrum because of the presence of numerous abundant nonphosphopeptides peaks from BSA. After enrichment with IMAC beads (Figure 9b), the mass spectrum was still dominated by the peaks derived from nonphosphopeptides, while only the peak at m/z 2061.76 can be assigned to phosphopeptide derived from β-casein. Similar results were also obtained by other researchers. Zou et al.32 investigated the
technical notes
Li et al.
Figure 8. MALDI mass spectra of peptide mixture containing peptides originating from tryptic digestions of 2 × 10-8 M of β-casein, ovalbumin, myoglobin, cytochrome C, and bovine serum albumin (BSA); (a) 0.5 µL, without enrichment; (b) on-bead analysis of phosphopeptides obtained when using Fe3O4@ZrO2 core-shell microspheres to selectively trap target peptides from the above tryptic digest mixture, 200 µL. Phosphopeptide ions derived from β-casein and ovalbumin are marked with asterisks. Phosphopeptide ions derived from R-casein are marked with pound sign. The metastable losses of phosphoric acid are indicated with 4. The data in parentheses are S/N of the corresponding peaks.
performance of Fe3+-IMAC to enrich phosphopeptides from the digest mixture of β-casein and BSA at different molar ratios; it was found that phosphopeptides could not be specifically captured when the molar ratio decreased to 1:10. Larsen et al.26 also demonstrated this using Fe3+-IMAC to enrich phosphopeptides from the digest mixture of phosphoproteins (R-casein, β-casein, ovalbumin) with nonphosphoproteins (serum albumin, β-lactoglobulin, carbonic anhydrase); when the ratio decreased to 1:50, the peaks of phosphopeptides could hardly be observed. These results indicated that Fe3+-IMAC lacks enough specificity to selectively capture phosphopeptides from a complex peptide mixture. However, when using our Fe3O4@ZrO2 core-shell microspheres, two monophosphopeptides of β-casein (marked with asterisks) and peaks derived from their metastable losses of phosphoric acid (marked with ∆) can be detected in the spectrum with high intensity, while nearly no nonphosphorylated peptides were detected (Figure 9c). The peak at m/z 1031.41 can be assigned to a double-charged ion of monophosphorylated peptide derived from β-casein. The peaks derived from R-casein (marked with pound signs) due to the impurity of β-casein used were also observed in the spectrum. The absence of the tetraphosphorylated peptide (at m/z 3122.3) may be due to the unique selectivity of ZrO2 for singly phosphorylated peptides.18 The results discussed above indicate a much more selective binding of the phosphorylated peptides on the Fe3O4@ZrO2 core-shell microspheres than on the commercial IMAC resin. Phosphorylated peptides present in low abundance in blood serum but play a vital role in regulatory mechanisms and may
serve as casual factor in diseases. The enrichment and analysis of phosphorylated peptides directly from human serum is a challenge for the researchers involved in phosphoproteomics. Hence, we also investigated the feasibility of employing our Fe3O4@ZrO2 core-shell microspheres for selective enrichment of phosphopeptides directly from dilution of human serum without any other pretreatment. Figure 10a presents the MALDI mass spectrum obtained using our Fe3O4@ZrO2 core-shell microspheres to selectively enrich target species from a dilution of human serum. Three remarkable peaks at m/z 1465.57, 1545.56, and 1616.60 appear in the mass spectrum. To confirm the identities of these peaks, MALDI-TOF-TOF mass spectrometry was employed for characterization. The tandem mass spectrum obtained by selecting the peptide ion at m/z 1465.57, 1545.56, and 1616.60 for low-energy CID are shown in Figure 10, panels b, c and d, respectively. All fragment ion spectra displayed a near-complete series of y-ion fragments. In Figure 10c,d, the mass difference of 167 Da between the peptide fragment ions y13 (1263.60) and y14 (1430.77) corresponds to a phosphoserine residue and located the phosphorylation site at Ser-3. Loss of phosphoric acid (H3PO4, 98 Da) from the precursor ion confirmed this assignment. These MS/MS ions were subjected to the NCBI protein database for matching of the possible proteins. The MS/MS results of the two peaks at m/z 1545.56 and 1616.60 are matched to two phosphopeptide residues derived from fibrinopeptide A (gi|229185, ADSGEGDFLAEGGGVR) whose increased level can be found in hepatocellular, ovarian, urothelial, and gastric cancer.33-36 The peak at m/z 1465.57 has the same sequence as the peak at m/z Journal of Proteome Research • Vol. 6, No. 11, 2007 4507
Fe3O4@ZrO2 Core-Shell Microspheres as Affinity Probes
technical notes
Figure 9. Comparison of the performance of Fe3O4@ZrO2 core-shell microspheres and IMAC beads for the selective enrichment of phosphorylated peptides from complex mixtures. Phosphorylated peptides from Peptide mixture 4 (β-casein and BSA with a molar ratio of 1:50) were enriched by IMAC (PHOS-select) or Fe3O4@ZrO2 core-shell microspheres. MALDI mass spectra obtained from without enrichment (a) and after enrichment with IMAC beads (b) and Fe3O4@ZrO2 core-shell microspheres (c). Phosphopeptide ions derived from β-casein are marked with asterisks. The metastable losses of phosphoric acid are indicated with ∆. The data in parentheses are S/N of the corresponding peaks.
1545.56, but without phosphorylation (Figure 10b). The results support that the Fe3O4@ZrO2 core-shell microspheres can be directly employed to enrich phosphopeptides from human serum without requiring any purification steps.
Conclusions In summary, we presented an innovational approach to synthesize Fe3O4@ZrO2 microspheres with well-defined coreshell structure, and the synthesized Fe3O4@ZrO2 core-shell microspheres were successfully applied for the enrichment and identification of phosphopeptides from protein tryptic digest mixture via a direct MALDI-TOF mass spectrometry analysis. The process of enrichment is very simple, quick, efficient, and specific. The resulting Fe3O4@ZrO2 core-shell, microspheresabsorbed phosphopeptides can be directly analyzed using 4508
Journal of Proteome Research • Vol. 6, No. 11, 2007
MALDI-TOF-MS analysis without elution form the Fe3O4@ZrO2 core-shell microspheres. In only 0.5 min, phosphopeptides sufficient for characterization by MALDI-MS can be enriched from very low concentrations of sample solutions. We also demonstrated that the Fe3O4@ZrO2 core-shell microspheres provide high trapping capacity with specificity and sensitivity for the enrichment of phosphorylated peptides from complex sample solutions, and the feasibility was further demonstrated by direct analysis of phosphopeptides in blood serum without any other pretreatment. All these results obtained indicate the Fe3O4@ZrO2 core-shell microspheres have high selectivity and sensitivity for enrichment of phosphorylated peptides, which opens up a possibility for their further application in analysis of real samples.
technical notes
Li et al.
Figure 10. (a) MALDI-TOF mass spectrum of phosphopeptides enriched by Fe3O4@ZrO2 core-shell microspheres from normal human serum. Phosphopeptide ions are marked with asterisks. MALDI-TOF/TOF mass spectra of the parent ions at (b) m/z 1465.57, (c) the parent ion at m/z 1545.56, and (d) the parent ion at m/z 1616.60 enriched by Fe3O4@ZrO2 core-shell microspheres. The amino acid sequence coverage is shown by yn ions. Journal of Proteome Research • Vol. 6, No. 11, 2007 4509
technical notes
Fe3O4@ZrO2 Core-Shell Microspheres as Affinity Probes
Acknowledgment. The work was supported by grants from 863 Project (No.2006AA02Z4C5), Shanghai Basic Research Priorities Programme (No.05dz19741), Natural Science Foundation of China (No.39870451), and Shanghai Municipal Commission for Science and Technology (No.0652nm006 and 0652nm018). References (1) Pandey, A.; Podtelejnikov, A. V.; Blagoev, B.; Bustelo, X. R.; Mann, M.; Lodish, H. F. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 179-184. (2) Steen, H.; Kuster, B.; Fernandez, M.; Pandey, A.; Mann, M. J. Biol. Chem. 2002, 277, 1031-1039. (3) Gronborg, M.; Kristiansen, T. Z.; Stensballe, A.; Andersen, J. S.; Ohara, O.; Mann, M.; Jensen, O. N.; Pandey, A. Mol. Cell. Proteomics 2002, 1, 517-527. (4) Ueda, E. K.; Gout, P. W.; Morganti, L. J. Chromatogr., A 2003, 988, 1-23. (5) Posewitz, M. C.; Tempst, P. Anal. Chem. 1999, 71, 2883-1892. (6) Raska, C. S.; Parker, C. E.; Dominski, A.; Marzluff, W. F.; Glish, G. L.; Pope, R. M.; Borchers, C. H. Anal. Chem. 2002, 74, 3429-3433. (7) Zhang, Y.; Yu, X.; Wang, X.; Shan, W.; Yang, P.; Tang, Y. Chem. Commun. 2004, 2882-2883. (8) Hirschberg, D.; Ja¨gerbrink, T.; Samskog, J.; Gustafsson, M.; Sta¨hlberg, M.; Alveliusm, G.; Husman, B.; Carlquist, M.; Jo¨rnvall, H.; Bergman, T. Anal. Chem. 2004, 76, 5864-5871. (9) Xu, X. Q.; Deng, C. H.; Gao, M. X.; Yu W. J.; Yang, P. Y.; Zhang, X. M. Adv. Mater. 2006, 18, 3289-3293. (10) Hart, S. R.; Waterfield, M. D.; Burlingame, A. L.; Cramer, R. J. Am. Soc. Mass Spectrom. 2002, 13, 1042-1051. (11) Chen, C. T.; Chen, Y. C. Anal. Chem. 2005, 77, 5912-5919. (12) Liang, S. S.; Makamba, H.; Huang, S. Y.; Chen, S. H. J. Chromatogr., A 2006, 1116, 38-45. (13) Rinalducci, S.; Larsen, M. R.; Mohammed, S.; Zolla, L. J. Proteome Res. 2006, 5, 973-982. (14) Coletti-Previero, M. A.; Previero, A. Anal. Biochem. 1989, 180, 1-10. (15) Koppel, R.; Litvak, M.; Solomon, B. J. Chromatogr., B 1994, 662, 191-196. (16) Wolschin, F.; Wienkoop, S.; Weckwerth, W. Proteomics 2005, 5, 4389-4397.
4510
Journal of Proteome Research • Vol. 6, No. 11, 2007
(17) Chen, C. T.; Chen, W. Y.; Tsai, P. J.; Chien, K. Y.; Yu, J. S.; Chen, Y. C. J. Proteome Res. 2007, 6, 316-325. (18) Kweon, H. K.; Håkansson, K. Anal. Chem. 2006, 78, 1743-1749. (19) Nandigala, P.; Chen, T. H.; Yang, C.; Hsu, W. H.; Heath, C. Biotechnol. Prog. 1997, 13, 844-848. (20) Deng, Y.; Deng, C.; Yang, D.; Wang, C.; Fu, S.; Zhang, X. Chem. Commun. 2005, 44, 5548-5550. (21) Gupta, P. K.; Hung, C. T. Life Sci. 1989, 44, 175-186. (22) Deng, Y.; Wang, C.; Shen, X.; Yang, W.; Jin, L.; Gao, H.; Fu, S. Chem.sEur. J. 2005, 11, 6006-6013. (23) Horak, D.; Rittich, B.; Safar, J.; Spanova, A.; Lenfeld, J.; Benes, M. J. Biotechnol. Prog. 2001, 17, 447-452. (24) Guiseppi-Elie, A.; Sheppar, N. F.; Brahim, S.; Narinesigh, D. Biotechnol. Bioeng. 2001, 75, 475-484. (25) Lo, C. Y.; Chen, W. Y.; Chen, C. T.; Chen, Y. C. J. Proteome Res. 2007, 6, 887-893. (26) Larsen, M. R.; Thingholm, T. E.; Jensen, O. N.; Roepstorff, P.; Jørgensen, T. J. D. Mol. Cell. Proteomics 2005, 4, 873-886. (27) Zhou, W.; Merrick, B. A.; Khaledi, M. G.; Tomer, K. B. J. Am. Soc. Mass Spectrom. 2000, 11, 273-282. (28) Amplett, C. B. Inorganic Ion Exchangers; Elsevier: Amsterdam, The Netherlands, 1964. (29) Blackwell, J. A.; Carr, P. W. J. Chromatogr., A 1991, 549, 43-57. (30) Blackwell, J. A.; Carr, P. W. J. Chromatogr., A 1991, 549, 59-75. (31) Muller, D. R.; Schindler, P.; Coulot, M.; Voshol, H.; van Oostrum, J. J. Mass Spectrom. 1999, 34, 336-345. (32) Zhou, H. J.; Xu, S. Y.; Ye, M. L.; Feng, S.; Pan, C.; Jiang, X. G.; Li, X.; Han, G. H.; Fu, Y.; Zou, H. F. J. Proteome Res. 2006, 5, 24312437. (33) Theodorescu, D.; Wittke, S.; Ross, M. M.; Walden, M.; Conaway, M.; Just, I.; Mischak, H.; Frierson, H. F. Lancet Oncol. 2006, 7, 230-40. (34) Bergen, H. R.; Vasmatzis, G.; Cliby, W. A.; Johnson, K. L.; Oberg, A. L.; Muddiman, D. C. Dis. Markers 2003, 19, 239-49. (35) Orvisky, E.; Drake, S. K.; Martin, B. M.; Abdel-Hamid, M.; Ressom, H. W.; Varghese, R. S.; An, Y.; Saha, D.; Hortin, G. L.; Loffredo, C. A.; Goldman, R. Proteomics 2006, 6, 2895-2902. (36) Ebert, M. P. A.; Niemeyer, D.; Deininger, S. O.; Wex, T.; Knippig, C.; Hoffmann, J.; Sauer, J.; Albrecht, W.; Malfertheiner, P.; Ro¨cken, C. J. Proteome Res. 2006, 5, 2152-2158.
PR070167S
