Phosphonate-modified metal oxides for the highly selective enrichment of phosphopeptides

Abstract

Metal oxide affinity chromatography has become one of the most widely used strategies for phosphopeptide enrichment prior to mass spectrometry analysis, but some defects still exist in this approach due to the complex physical and chemical properties of the metal oxide surface. Although the simultaneous phosphorylation of adjacent amino acids may greatly affect the bioactivity of the protein, there are few reports on the specific enrichment of multi-phosphopeptides. In this work, we report a highly selective enrichment method for capturing phosphopeptides or multi-phosphopeptides based on phosphonate-modified metal oxides. Compounds with different numbers of phosphate groups were adsorbed on the surface of ZrO₂ and TiO₂ to obtain phosphonate-modified metal oxides. Among them, phosphoric acid modified metal oxides (1P-ZrO₂ and 1P-TiO₂) could significantly enhance their selectivity towards phosphopeptides; and alendronate-modified metal oxides (2P-ZrO₂ and 2P-TiO₂) showed high selectivity for the enrichment of multi-phosphopeptides. In addition, the detection sensitivity was greatly improved by using these novel materials. The mechanism of the specific enrichment was considered to be ligand exchange and blocking of strong adsorption sites by the compounds containing the phosphate group. Finally, tryptic digests of proteins of human Jurkat-T cell lysate were further used to demonstrate the selectivity and specificity of ZrO₂, 1P-ZrO₂ and 2P-ZrO₂.

---

1. Introduction

Reversible phosphorylation of proteins, one of the most important post-translational modifications, plays a crucial role in a large number of biological processes, including signal transduction, cell proliferation, differentiation and apoptosis.^1–3 In particular, the simultaneous phosphorylation of adjacent amino acids is considered to greatly affect the regulatory activity of the protein.^4,5 Currently, mass spectrometry (MS) based techniques have become the first and foremost choice for the analysis of protein phosphorylation because of their high sensitivity and accuracy.^6–8 However, analysis of phosphopeptides by MS is often hampered by their low abundance, sub-stoichiometry and lower ionization efficiency than non-phosphopeptides. Therefore, selective and sensitive enrichment methods are essential for global phosphoproteomic analysis.^9,10

Various methods for phosphopeptide enrichment have been widely explored, such as antibody-based affinity enrichment, immobilized metal affinity chromatography (IMAC) and metal oxide affinity chromatography (MOAC).^11–14 Among them, MOAC is one of the most powerful and extensively used enrichment methods, which is based on the affinity between metal oxides and phosphate group.¹⁵ Although various metal oxides have been widely explored for phosphopeptide enrichment, their selectivity is greatly different due to the complex physical and chemical properties of metal oxide surface.^16–20 For example, compared with TiO₂, ZrO₂ showed higher selectivity for mono-phosphopeptides, and more nonspecific binding of non-phosphopeptides when used for large-scale studies.^15,21 The difference was attributed to the fact that ZrO₂ is a stronger Lewis acid than TiO₂, and the coordination numbers of Zr and Ti are different in the crystalline forms, which result in different binding properties for mono- and multi-phosphopeptides.²¹ The different selectivity not only exists in different metal oxides, but also for the same metal oxide synthesized by different methods.^22,23 It has been reported that rutile-form titania exhibited higher selectivity in phosphopeptide enrichment than commercial titania.²² Additionally, Leitner et al. compared nanocast TiO₂ spheres with commercial TiO₂ material, and found that both TiO₂ materials yielded the successful identification of a comparable total number of phosphopeptides, but the overlap of the two data sets was less than one-third.²³ These differences might be due to the different surface chemistry of them, but the exact mechanism remains to be explored. Therefore, research of the properties of metal oxide surface is significant for understanding the interaction of metal oxide and phosphopeptides.

On the other hand, most of the reported methods take aim at capturing both of mono- and multi-phosphorylated peptides. However, when the co-eluting mono- and multi-phosphorylated peptides are simultaneously analyzed by MS, the detection of multi-phosphopeptides is frequently hampered by the ion suppression effect due to their lower ionization efficiency than mono-phosphorylated peptides.^24,25 Therefore, methods which can effectively isolate and enrich multi-phosphopeptides from mono-phosphopeptides and complex biological samples will be beneficial to the sensitive determination of multi-phosphopeptides. To date, only a few studies have been reported for the specific enrichment of multi-phosphopeptides.^26–30 However, these methods are restricted to expensive reagents and complex synthetic procedures of materials. Therefore, a simple and efficient method for the enrichment and separation of multi-phosphopeptides is still very attractive.

In this study, to address above issues, we present a novel strategy for highly selective enrichment of phosphopeptides or multi-phosphopeptides based on phosphonate-modified metal oxides. A series of phosphonate-modified metal oxides were prepared by adsorbing compounds with different numbers of phosphate group on the surface of ZrO₂ and TiO₂. The phosphonate-modified metal oxides displayed different selectivity towards phosphopeptides and multi-phosphopeptides. The selectivity and detection sensitivity of the modified metal oxides were greatly improved. And then, the possible mechanism of the specific enrichment of phosphopeptides or multi-phosphopeptides by phosphonate-modified metal oxides was proposed. Finally, phosphonate-modified metal oxides were further applied for the enrichment of phosphopeptides from tryptic digests of proteins of Jurkat-T cell lysate.

2. Materials and methods

2.1. Chemicals and materials

Alendronate sodium trihydrate, nitrilotri(methylphosphonic acid) (ATMP), and diethylenetriaminepentakis(methylphosphonic acid) solution (DTPMP) were purchased from Aladdin Chemical Reagent Co. (Shanghai, China). HPLC grade acetonitrile (ACN) was purchased from Fisher Scientific (Pittsburgh, PA, USA). Titanium(IV) oxide (TiO₂, anatase), zirconium(IV) oxide (ZrO₂, powder), phosphoric acid (H₃PO₄), trifluoroacetic acid (TFA), 2,5-dihydroxybenzoic acid (2,5-DHB), bovine α-casein, bovine β-casein, bovine serum albumin (BSA), dithiothreitol (DTT), iodoacetamide (IAA) and trypsin were purchased from Sigma-Aldrich (St. Louis, MO, USA). Tris(hydroxymethyl)aminomethane (Tris), urea, acetone, ethanol, acetic acid and calcium chloride (CaCl₂) were all of analytical reagent grade and supplied by Shanghai General Chemical Reagent Factory (Shanghai, China). Purified water was obtained with a Milli-Q apparatus (Millipore, Bedford, MA, USA).

2.2. Preparation of phosphonate-modified metal oxides

Thirty milligram of ZrO₂ was dispersed in 1 mL of alendronate sodium solution (0.01 M). And the mixture was incubated at 60 °C for 1 h. After washing with distilled water for several times, the obtained alendronate-modified ZrO₂ particles which were denoted as 2P-ZrO₂, were re-suspended in 1 mL of distilled water. H₃PO₄-ZrO₂ (1P-ZrO₂), ATMP-ZrO₂ (3P-ZrO₂), DTPMP-ZrO₂ (5P-ZrO₂), H₃PO₄-TiO₂ (1P-TiO₂), alendronate-TiO₂ (2P-TiO₂), ATMP-TiO₂ (3P-TiO₂), and DTPMP-TiO₂ (5P-TiO₂) were prepared using the same procedure. For clarity, the chemical structures of H₃PO₄, alendronic acid, ATMP and DTPMP are shown in Fig. 1.

Fig. 1 The chemical structures of phosphoric acid, alendronic acid, ATMP and DTPMP.

2.3. Sample pretreatment

Bovine α-casein and β-casein was originally made up into stock solutions at 1 mg mL⁻¹ with Milli-Q water, respectively. Proteins were digested with trypsin by using an enzyme to substrate ratio of 1 : 50 (w/w) in 100 mM Tris–HCl (pH 8.5) at 37 °C for 16 h.

BSA (1 mg) was dissolved in 100 μL of denaturing buffer solution (8 M urea in 100 mM Tris–HCl pH 8.5). The dissolved BSA was reduced by 10 mM DTT for 30 min at 37 °C and alkylated by 20 mM IAA for 30 min at room temperature in the dark. The reduced and alkylated protein mixture was diluted with 300 μL 100 mM Tris–HCl (pH 8.5), and then digested with trypsin at an enzyme to substrate ratio of 1 : 50 (w/w) by incubating at 37 °C for 16 h. All the tryptic digests were lyophilized to dryness and stored at −80 °C before use.

Human Jurkat-T cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin at 37 °C in a humidified atmosphere containing 5% CO₂. The cells were harvested and lysed in RIPA lysis buffer (1% NP-40, 0.25% deoxycholate) with protease and phosphatase inhibitors on ice for 30 min. The cell lysate was centrifuged at 16 000g under 4 °C for 30 min and the protein concentration of the supernatant was determined by bicinchoninic acid assay. Proteins were precipitated with 1.5 mL of 50% acetone–50% ethanol–0.1% acetic acid on ice for 1 h, and then centrifuged at 1800g under 4 °C for 30 min. The protein pellet was re-suspended in 2 mL of buffer containing 8 M urea, 0.2 M Tris (pH 8.0), and 4 mM CaCl₂. The following digestion procedure was the same as that of BSA, and the digested products were desalted by C18 cartridge and stored at −20 °C before use.

2.4. Phosphopeptide enrichment procedure

For phosphopeptide enrichment from tryptic digests of standard proteins, the suspension of phosphonate-modified metal oxides (5 μL 30 mg mL⁻¹) was added to 20 μL of peptide mixture (containing 1 pmol of tryptic digest of α-casein) and vortexed for 30 min. After washing with 30 μL of 1% TFA–50% ACN twice, the trapped peptides were eluted with 30 μL 2.5% ammonium hydroxide. The eluted solution was then lyophilized to dryness. Two microliter of matrix solution (mixture of 20 mg mL⁻¹ 2,5-DHB in 50% (v/v) ACN, 1% (v/v) phosphoric acid) was introduced into the elute and 1 μL of the mixture was used for MALDI-TOF MS analysis.

For phosphopeptide enrichment from tryptic digests of proteins of Jurkat-T cell lysate, 50 mg ZrO₂, 1P-ZrO₂ or 2P-ZrO₂ were mixed with 1 mg tryptic digests in 1 mL 1% TFA–50% ACN and incubated at 37 °C for 45 min. After washing with 1 mL 1% TFA–50% ACN twice, the trapped peptides were eluted with 1 mL 2.5% ammonium hydroxide. The eluted solution was then lyophilized to dryness, desalted with Ziptip C18 and used for RPLC-ESI-MS/MS analysis.

2.5. Mass spectrometry analysis

All MALDI-TOF MS spectra were recorded with an Axima TOF² mass spectrometry (Shimadzu, Kyoto, Japan). The instrument was equipped with a 337 nm nitrogen laser with a 3 ns pulse width. The detection was performed in positive ion reflector mode with an accelerating voltage of 20 kV. Typically, 500 laser shots were averaged to generate each spectrum.

RPLC-ESI-MS/MS was used to analyze the sample from Jurkat-T cells. The analysis was carried out on a hybrid quadrupole-TOF LC-MS/MS mass spectrometer (TripleTOF 5600+, ABSciex) equipped with a nanospray source. Peptides were first loaded onto a C18 trap column (5 mm × 0.3 mm i.d., 5 μm, Agilent Technologies) and then eluted into a C18 analytical column (150 mm × 75 μm i.d., 3 μm, 100 Å, Eksigent). Mobile phase A (3% DMSO, 97% H₂O, 0.1% formic acid) and mobile phase B (3% DMSO, 97% ACN, 0.1% formic acid) were used to establish a 100 min gradient, which was comprised of 0 min in 5% B, 65 min of 5–23% B, 20 min of 23–52% B, 1 min of 52–80% B, 4 min of 80% B, 0.1 min of 80–5% B, and 10 min of 5% B. A constant flow rate was set at 300 nL min⁻¹. MS scans were conducted from 350 to 1500 amu, with a 250 ms time span. For MS/MS analysis, each scan cycle consisted of one full-scan mass spectrum (with m/z ranging from 350 to 1500 and charge states from 2 to 5) followed by 40 MS/MS events. The threshold count was set to 120 to activate MS/MS accumulation and former target ion exclusion was set for 18 s.

2.6. Data analysis

Raw data from TripleTOF 5600+ were analyzed with Protein Pilot Software 4.5. Data were searched against the Uniprot human reference proteome database (version 201310) using the following parameters: sample type, identification; cys alkylation, iodoacetamide; digestion, trypsin; special factors, phosphorylation emphasis. Search effort was set to rapid ID. A 1% critical false discovery rates in Protein Pilot was selected to calculate the number of identifications.

3. Results and discussion

3.1. Comparison of different modifiers

Various modifiers containing different numbers of phosphate group (Fig. 1) can be easily adsorbed on the surface of ZrO₂ and TiO₂ based on the affinity of phosphate group to metal oxide.^31–35 Therefore, a series of phosphonate-modified metal oxides were prepared in this work (Table 1). The selectivity of different phosphonate-modified metal oxides for phosphopeptide enrichment was investigated by using tryptic digest of α-casein (Fig. 2 and 3). As shown in Fig. 2, ZrO₂, 1P-ZrO₂, 2P-ZrO₂, 3P-ZrO₂ and 5P-ZrO₂ displayed different selectivity towards phosphopeptides. After enrichment with ZrO₂, twelve phosphopeptides (five mono-phosphopeptides and seven multi-phosphopeptides) can be observed in Fig. 2(a). The sequences of the observed phosphopeptides are listed in Table 2. However, after enrichment with 1P-ZrO₂, fifteen phosphopeptides (five mono-phosphopeptides and ten multi-phosphopeptides) are observed in Fig. 2(b), indicating that the specificity of 1P-ZrO₂ towards phosphopeptides is better than ZrO₂, especially for multi-phosphopeptides. While using 2P-ZrO₂ for the enrichment, eleven phosphopeptides (one mono-phosphopeptides and ten multi-phosphopeptides) are detected in Fig. 2(c). Comparing with 1P-ZrO₂, 2P-ZrO₂ displayed predominant selectivity for multi-phosphopeptides. As for 3P-ZrO₂, although there are also eleven phosphopeptides (the same as the ones obtained by 2P-ZrO₂) which can be detected in Fig. 2(d), the signal-to-noise ratio (SNR) of mass spectrometric peaks was worse than that obtained by 2P-ZrO₂ (Table 3). At last, none of peaks can be detected in Fig. 2(e), indicating that 5P-ZrO₂ exhibited no enrichment ability for phosphopeptides.

Table 1 The materials prepared in this work

Metal oxide	Ligand
	Phosphoric acid	Alendronic acid	ATMP	DTPMP
ZrO2	1P-ZrO2	2P-ZrO2	3P-ZrO2	5P-ZrO2
TiO2	1P-TiO2	2P-TiO2	3P-TiO2	5P-TiO2

---

Fig. 2 MALDI mass spectra of tryptic digest of α-casein enriched with ZrO₂ (a), 1P-ZrO₂ (b), 2P-ZrO₂ (c), 3P-ZrO₂ (d), or 5P-ZrO₂ (e). The concentration α-casein is 5 × 10⁻⁸ M (1 pmol). The superscript represents the number of phosphorylation site.

Fig. 3 MALDI mass spectra of tryptic digests of α-casein enriched with TiO₂ (a), 1P-TiO₂ (b), 2P-TiO₂ (c), 3P-TiO₂ (d), or 5P-TiO₂ (e). The concentration α-casein is 5 × 10⁻⁸ M (1 pmol). The superscript represents the number of phosphorylation site.

Table 2 Detailed information of the observed phosphopeptides obtained from tryptic digests of α-casein^a

No.	[M + H]^+	Phosphorylation site	Amino acid sequence	ZrO2	1P-ZrO2	2P-ZrO2	TiO2	1P-TiO2	2P-TiO2
a “S̲” and “M*” represents phosphorylation on serine and oxidation on methionine respectively.
α1	1466.6	1	TVDMES̲TEVFTK	+	+	−	−	−	−
α2	1539.7	2	EQLS̲TS̲EENSKK	+	+	+	−	+	+
α3	1660.8	1	VPQLEIVPNS̲AEER	+	+	−	+	+	−
α4	1847.7	1	DIGSES̲TEDQAMEDIK	+	+	−	+	+	−
α5	1927.7	2	DIGS̲ES̲TEDQAMEDIK	+	+	+	+	+	+
α6	1943.7	2	DIGS̲ES̲TEDQAM*EDIK	+	+	+	+	+	+
α7	1952.0	1	YKVPQLEIVPNS̲AEER	+	+	−	+	+	+
α8	2619.0	4	NTMEHVS̲S̲S̲EES̲IISQETYK	−	+	+	+	+	+
α9	2678.0	3	VNELS̲KDIGS̲ES̲TEDQAMEDIK	+	+	+	+	+	+
α10	2703.5	1	LRLKKYKVPQLEIVPNS̲AEERL	+	+	+	+	+	+
α11	2720.9	5	QMEAES̲IS̲S̲S̲EEIVPNS̲VEQK	+	+	+	+	+	+
α12	2736.9	5	QM*EAES̲IS̲S̲S̲EEIVPNS̲VEQK	−	+	+	+	+	+
α13	2935.1	3	EKVNELS̲KDIGS̲ES̲TEDQAMEDIKQ	+	+	+	+	+	+
α14	3008.0	4	NANEEEYSIGS̲S̲S̲EES̲AEVATEEVK	+	+	+	+	+	+
α15	3088.0	5	NANEEEYS̲IGS̲S̲S̲EES̲AEVATEEVK	−	+	+	+	+	+

---

Table 3 The SNR of α-casein phosphopeptides enriched by 2P-ZrO₂, 3P-ZrO₂, 2P-TiO₂ or 3P-TiO₂

Materials
Phosphopeptides	SNR
	2P-ZrO2	3P-ZrO2	2P-TiO2	3P-TiO2
a “—” represents phosphopeptide undetected.
α2	7.00	5.32	12.83	5.12
α5	123.61	37.60	91.67	10.71
α6	38.24	5.19	127.33	13.74
α7	—^a	—	23.17	6.72
α8	11.10	9.71	4.33	4.31
α9	7.58	5.70	6.99	4.07
α10	20.78	19.34	11.81	10.27
α11	16.84	24.26	22.08	26.61
α12	6.34	6.25	15.74	13.98
α13	12.23	11.84	6.66	4.14
α14	14.43	18.17	17.44	9.76
α15	6.18	7.10	6.06	4.65

---

Similarly, TiO₂, 1P-TiO₂, 2P-TiO₂, 3P-TiO₂ and 5P-TiO₂ also displayed different selectivity for the enrichment of phosphopeptides (Fig. 3). Thirteen (four mono-phosphopeptides and nine multi-phosphopeptides) and fourteen phosphopeptides (four mono-phosphopeptides and ten multi-phosphopeptides) can be detected after enrichment with TiO₂ and 1P-TiO₂, respectively (Fig. 3(a) and (b)). Therefore, the selectivity of 1P-TiO₂ is slightly better than TiO₂. Twelve phosphopeptides (two mono-phosphopeptides and ten multi-phosphopeptides) are detected in Fig. 3(c), indicating that 2P-TiO₂ also has advantage for capturing multi-phosphopeptides. Like 3P-ZrO₂ and 5P-ZrO₂, although twelve phosphopeptides can be detected with 3P-TiO₂, the SNR of the observed peptides was worse than that enriched with 2P-TiO₂, and none of phosphopeptides can be detected after enrichment with 5P-TiO₂ (Table 3, Fig. 3(d) and (e)). Through the contrast of nP-ZrO₂ and nP-TiO₂ (n refers to 0, 1, 2, 3, 5), ZrO₂ showed better selectivity for mono-phosphopeptides than TiO₂, whereas TiO₂ preferentially enriched multi-phosphopeptides as also described in the previous report.²¹ However, the specificity of 1P-ZrO₂ is nearly the same with 1P-TiO₂. And 2P-ZrO₂ exhibited better selectivity for multi-phosphopeptides than 2P-TiO₂.

3.2. Investigation of the selectivity

In order to further compare the selectivity toward phosphopeptides, ZrO₂, 1P-ZrO₂, 2P-ZrO₂, TiO₂, 1P-TiO₂ and 2P-TiO₂ were used to enrich phosphopeptides from tryptic digest mixtures of β-casein and BSA with different molar ratios (β-casein: BSA = 1 : 10, 1 : 50 and 1 : 200). When ZrO₂ was used for the enrichment of tryptic digests of β-casein and BSA, although three or two phosphopeptides derived from β-casein can be observed, the peaks of non-phosphopeptides dominated the mass spectra (Fig. 4(a)–(c)). Moreover, the signal intensities of phosphopeptides decreased gradually with increasing concentration of BSA. The sequences of the observed phosphopeptides are listed in Table 4. However, after enrichment with 1P-ZrO₂, clear mass spectra of four phosphopeptides can be observed in Fig. 4(d)–(f), and the intensive peaks of non-phosphopeptides disappeared, indicating the better specificity of 1P-ZrO₂ than ZrO₂. When the tryptic digests of β-casein and BSA were enriched with 2P-ZrO₂, two multi-phosphopeptides can be easily detected with strong intensity and freed from the interference of non-phosphopeptides and mono-phosphopeptides, even if the molar ratios of β-casein over BSA were decreased to 1 : 200 (Fig. 4(g)–(i)). The above results demonstrate that 1P-ZrO₂ and 2P-ZrO₂ possess excellent capability for the selective enrichment of phosphopeptides and multi-phosphopeptides from a complex peptide mixture, respectively.

Fig. 4 MALDI mass spectra of the tryptic digest mixture of β-casein and BSA enriched with ZrO₂ (a–c), 1P-ZrO₂ (d–f) and 2P-ZrO₂ (g–i). Molar ratios of β-casein to BSA are 1 : 10 (a, d and g), 1 : 50 (b, e and h) and 1 : 200 (c, f and i). The superscript represents the number of phosphorylation site.

Table 4 Detailed information of the observed phosphopeptides obtained from tryptic digests of β-casein

No.	[M + H]^+	Phosphorylation site	Sequence
β1	2061.8	1	FQS̲EEQQQTEDELQDK
β2	2556.7	1	FQS̲EEQQQTEDELQDKIHPF
β3	2966.2	4	ELEELNVPGEIVES̲LS̲S̲S̲EESITR
β4	3122.3	4	RELEELNVPGEIVES̲LS̲S̲S̲EESITR

---

Similar results can also be seen from the case of TiO₂. Fig. 5 showed the mass spectra of tryptic digests of β-casein and BSA after enrichment with TiO₂, 1P-TiO₂ and 2P-TiO₂. As shown in Fig. 5(a)–(c), the mass spectra of phosphopeptides after enrichment with TiO₂ are also interfered with non-phosphopeptides, although the interference was less than that obtained with ZrO₂. However, after enrichment with 1P-TiO₂, the mass spectra were cleaner than that obtained with TiO₂, just as the case of 1P-ZrO₂ (Fig. 5(d)–(f)). When 2P-TiO₂ was used for the enrichment, two multi-phosphopeptides (β3 and β4) were also detected without the interference of non-phosphopeptides and mono-phosphopeptides (Fig. 5(g)–(i)). By comparing 2P-ZrO₂ and 2P-TiO₂, we can see that 2P-ZrO₂ exhibited better selectivity for multi-phosphopeptides, because there is a weak signal of β1 detected in Fig. 5(g) when the molar ratio of β-casein over BSA was 1 : 10. To summarize, through the above comparisons, we can see that both 1P-ZrO₂ and 1P-TiO₂ possess excellent capability for the selective enrichment of phosphopeptides from a complex peptide mixture, while ZrO₂ and TiO₂ were subjected to the interference of abundant non-phosphopeptides. And 2P-ZrO₂ and 2P-TiO₂ showed high selectivity for the enrichment of multi-phosphopeptides, while the selectivity of 2P-ZrO₂ was better than that of 2P-TiO₂.

Fig. 5 MALDI mass spectra of the tryptic digest mixture of β-casein and BSA enriched with TiO₂ (a–c), 1P-TiO₂ (d–f) and 2P-TiO₂ (g–i). Molar ratios of β-casein to BSA are 1 : 10 (a, d and g), 1 : 50 (b, e and h) and 1 : 200 (c, f and i). The superscript represents the number of phosphorylation site.

3.3. Comparison of detection sensitivity

The detection sensitivity of phosphopeptides by ZrO₂, 1P-ZrO₂, 2P-ZrO₂, TiO₂, 1P-TiO₂ and 2P-TiO₂ was investigated by tryptic digest of β-casein with different amounts (200, 50, and 20 fmol) (Fig. 6 and 7). ZrO₂ was first applied for the enrichment of tryptic digest of β-casein. When 200 fmol β-casein digest was used, three phosphopeptides can be detected (Fig. 6(a)). However, when the amount of β-casein digest was 50 fmol, only one phosphopeptide (β1) can be observed (Fig. 6(b)). If the amount of β-casein digest was further decreased to 20 fmol, a weak signal of β1 with the SNR of 5.15 can be observed (Fig. 6(c)). For 1P-ZrO₂, three and two phosphopeptides can be detected, respectively, when the amount of β-casein digest was 200 fmol and 50 fmol (Fig. 6(d) and (e)). Moreover, even if the amount of β-casein digest was decreased to 20 fmol, two phosphopeptides (β1 with the SNR of 11.55 and β4 with the SNR of 4.26) can be detected (Fig. 6(f)). Therefore, the sensitivity of 1P-ZrO₂ for capturing phosphopeptides was better than that of ZrO₂. 2P-ZrO₂ was also used for the enrichment of tryptic digest of β-casein. As shown in Fig. 6(g)–(i), only the multi-phosphopeptides β4 can be observed, and the SNR of β4 was 10.05 when the amount of β-casein digest was 20 fmol, indicating that 2P-ZrO₂ can provide better sensitivity for capturing multi-phosphopeptides than 1P-ZrO₂.

Fig. 6 MALDI mass spectra of tryptic digest of β-casein with different amounts enriched with ZrO₂ (a–c), 1P-ZrO₂ (d–f) and 2P-ZrO₂ (g–i). The amounts of β-casein are 100 (a, d and g), 20 (b, e and h), and 10 fmol (c, f and i). The superscript represents the number of phosphorylation site.

Fig. 7 MALDI mass spectra of tryptic digest of β-casein with different amounts enriched with TiO₂ (a–c), 1P-TiO₂ (d–f) and 2P-TiO₂ (g–i). The amounts of β-casein are 100 (a, d and g), 20 (b, e and h), and 10 fmol (c, f and i). The superscript represents the number of phosphorylation site.

While using TiO₂ for the enrichment of tryptic digests of β-casein with the amount of 200 and 50 fmol, three and two phosphopeptides can be detected, respectively (Fig. 7(a) and (b)). When the amount of β-casein digest was decreased to 20 fmol, two phosphopeptides (β1 with the SNR of 9.80 and β4 with the SNR of 4.14) can still be detected (Fig. 7(c)). Comparing with ZrO₂, TiO₂ possesses better sensitivity for capturing phosphopeptides. For 1P-TiO₂, three and two phosphopeptides can also be detected in Fig. 7(d) and (e), and the two phosphopeptides (β1 with the SNR of 11.60 and β4 with the SNR of 4.27) can still be detected in Fig. 7(f). The specificity of 1P-TiO₂ is only slightly better than TiO₂, and similar to that of 1P-ZrO₂. The multi-phosphopeptides β4 can be easily detected with 2P-TiO₂ (Fig. 7(g)–(i)). And the SNR of β4 was 10.25 when the amount of β-casein digest was 20 fmol, which is similar to the result obtained by 2P-ZrO₂. However, there is still one mono-phosphopeptide β1 observed in Fig. 7(g), indicating that the specificity of 2P-TiO₂ for multi-phosphopeptides is worse than that of 2P-ZrO₂, which was consistent with the above results. Therefore, from Fig. 6 and 7, we can see that 1P-ZrO₂ and 1P-TiO₂ possess better sensitivity for the selective enrichment of phosphopeptides than ZrO₂ and TiO₂. And 2P-ZrO₂ and 2P-TiO₂ showed the best specificity for multi-phosphopeptides.

3.4. The enrichment mechanism

According to the above results, we proposed the possible mechanism of the specific enrichment of phosphopeptides by the phosphonate-modified metal oxides. First, phosphopeptides were captured by ligand exchange. For 1P-ZrO₂ and 1P-TiO₂, the adsorption sites on the surface of metal oxides were first occupied by phosphate groups, and phosphopeptides were captured by ligand exchange with the phosphate groups adsorbed on 1P-ZrO₂ or 1P-TiO₂. Because the lower affinity of carboxylate group than phosphate group, non-phosphopeptides cannot displace phosphate group. Therefore, 1P-ZrO₂ and 1P-TiO₂ will exhibit better selectivity of phosphopeptides and less nonspecific binding of non-phosphopeptides than ZrO₂ and TiO₂. For 2P-ZrO₂ and 2P-TiO₂, ZrO₂ and TiO₂ were first modified by alendronate which has two phosphate groups. As a result, multi-phosphopeptides can be captured by ligand exchange with alendronate. Besides, the mono-phosphopeptides can't replace alendronate and they are barely enriched by 2P-ZrO₂ and 2P-TiO₂. By this enrichment process, multi-phosphopeptides can be separated from the non- and mono-phosphopeptides and apart from their interference prior to MS analysis. For 3P-ZrO₂ and 3P-TiO₂, multi-phosphopeptides have been hard to replace ATMP which has three phosphate groups. Therefore, low recovery was obtained. While for 5P-ZrO₂ and 5P-TiO₂, DTPMP which has five phosphate groups is hard to be replaced due to its stronger interaction with metal oxides. Therefore, none of phosphopeptides can be captured.

In addition, 1P-ZrO₂ and 1P-TiO₂ showed higher recovery towards phosphopeptides (especially for multi-phosphopeptides) than ZrO₂ and TiO₂, which might be ascribed to different surface geometrical structure and morphology on the surface of metal oxides.^36–40 For example, Morterra et al. studied the adsorption of carbon dioxide on the surface of ZrO₂, and they found that there are three kinds of Lewis acid centers on the surface of ZrO₂ which can be capable of linearly chemisorbing CO₂.³⁷ Auroux and Gervasini stated that there were two kinds of strong adsorption Lewis acid sites on the surface of ZrO₂ by adsorbing ammonia.³⁸ The heterogeneity of TiO₂ surface was also reported by Hadjiivanov et al.⁴⁰ According to their research, there were three kinds of Lewis acid sites (Ti⁴⁺ sites) on the surface of TiO₂. All of these reports demonstrated that the surface of metal oxides is heterogeneous. Therefore, Lewis acid sites of different strengths existing on the surface of metal oxides were denoted as weak and strong adsorption sites, respectively, in the current work. While using ZrO₂ and TiO₂ for capturing phosphopeptides, all adsorption sites on their surface can be applied for phosphopeptides. In this case, the phosphopeptides adsorbed on the strong adsorption sites are hard to be desorbed, especially for multi-phosphopeptides, therefore low recovery is obtained. However, as for 1P-ZrO₂ and 1P-TiO₂, all of adsorption sites are first occupied by phosphate groups. While using 1P-ZrO₂ and 1P-TiO₂ for phosphopeptide enrichment, the phosphate groups adsorbed on the weak adsorption sites will be replaced by phosphopeptides, and the ones adsorbed on the strong adsorption sites will retain. Therefore, phosphopeptides can be captured by 1P-ZrO₂ and 1P-TiO₂ through ligand exchange mechanism and high recovery can be obtained. Similarly, while using 2P-ZrO₂ and 2P-TiO₂ for multi-phosphopeptide enrichment, high recovery can be obtained due to the strong adsorption sites have been occupied by alendronate. Fig. 8 is the schematic diagram of the possible extraction mechanism by phosphonate-modified metal oxides.

Fig. 8 The possible mechanism for the selective enrichment of phosphopeptide or multi-phosphopeptide using phosphonate-modified metal oxides.

3.5. Phosphopeptide enrichment from tryptic digests of proteins of cell lysate

Encouraged by the high specificity of phosphonate-modified metal oxides, ZrO₂, 1P-ZrO₂ and 2P-ZrO₂ were further used to enrich phosphopeptides from tryptic digests of proteins of Jurkat-T cell lysate. A total of 1 mg tryptic digests of proteins of Jurkat-T cell lysate were treated with ZrO₂, 1P-ZrO₂ or 2P-ZrO₂, respectively, and the elutes were analyzed by RPLC-ESI-MS/MS. In total, 6767 unique peptides including 1002 phosphopeptides were identified after enrichment with ZrO₂, while 2962 unique peptides including 2237 phosphopeptides were identified after enrichment with 1P-ZrO₂ (see the supporting material peptide summary†). Phosphopeptide selectivity (the ratio of the number of phosphopeptides to that of all identified peptides) was calculated to be 14.8% for ZrO₂, and 75.5% for 1P-ZrO₂. According to the contrast, we can see that 1P-ZrO₂ displayed higher specificity for phosphopeptides than ZrO₂. The number of unique peptides captured by 1P-ZrO₂ greatly decreases while compared with ZrO₂, which is because the adsorption sites of 1P-ZrO₂ were occupied by phosphate groups.

In addition, 51 unique peptides including 43 phosphopeptides (nine mono-phosphopeptides and thirty-four multi-phosphopeptides) were identified after enrichment with 2P-ZrO₂. The number of phosphopeptides identified by 2P-ZrO₂ was relatively few, which can be attributed to the following reasons. The adsorption sites of 2P-ZrO₂ were occupied by alendronate; and the complexity of the sample was also a very significant reason. However, phosphopeptide selectivity was calculated to be 84.3% for 2P-ZrO₂. Among the identified phosphopeptides, 79.1% of which were multi-phosphopeptides. The results suggest that nonspecific binding of non-phosphopeptides were few for 2P-ZrO₂, and the predominant peptides enriched by 2P-ZrO₂ were multi-phosphopeptides. Although the extracting efficiency of 2P-ZrO₂ in the product of cell lysate were worse than that in less complex peptide mixtures, the selectivity for multi-phosphopeptides is still higher than mono-phosphopeptides.

4. Conclusion

In this study, phosphonate-modified metal oxides (nP-ZrO₂ and nP-TiO₂, n refers to 0, 1, 2, 3, 5) were prepared and applied for the enrichment of phosphopeptides or multi-phosphopeptide. Metal oxides modified with compounds containing different numbers of phosphate group displayed different selectivity for phosphopeptides and multi-phosphopeptides. Compared with ZrO₂ and TiO₂, 1P-ZrO₂ and 1P-TiO₂ showed higher selectivity and specificity for the enrichment of phosphopeptides. And, 2P-ZrO₂ and 2P-TiO₂ showed highly selectivity and specificity for the enrichment of multi-phosphopeptides. While comparing different metal oxides, although ZrO₂ showed poor extraction efficiency than TiO₂, the selectivity and specificity of 1P-ZrO₂ and 1P-TiO₂ are similar, and 2P-ZrO₂ displayed better specificity than 2P-TiO₂. Furthermore, we proposed the selective enrichment mechanism towards phosphopeptides and multi-phosphopeptide by phosphonate-modified metal oxides, which can be ascribed to be ligand exchange and blocking of strong adsorption sites by phosphate group-containing compounds. Finally, for tryptic digests of proteins of human Jurkat-T cell lysate, 1P-ZrO₂ also displayed higher specificity for phosphopeptides than ZrO₂, and 2P-ZrO₂ displayed high specificity for multi-phosphopeptides. Our work will promote the research of MOAC mechanism and its application in phosphopeptide enrichment, and provide a new simple method for capturing multi-phosphopeptides.

Acknowledgements

The authors thank the financial support from the National Basic Research Program of China (973 Program) (2013CB910702).

References

1. T. Hunter, Cell, 2000, 100, 113–127.
2. S. A. Johnson and T. Hunter, Nat. Biotechnol., 2004, 22, 1093–1094.
3. J. V. Olsen, B. Blagoev, F. Gnad, B. Macek, C. Kumar, P. Mortensen and M. Mann, Cell, 2006, 127, 635–648.
4. S. M. Plafker, A. S. Woods and W. Gibson, J. Virol., 1999, 73, 9053–9062.
5. P. Cohen, Trends Biochem. Sci., 2000, 25, 596–601.
6. D. T. McLachlin and B. T. Chait, Curr. Opin. Chem. Biol., 2001, 5, 591–602.
7. M. Mann, S. E. Ong, M. Gronborg, H. Steen, O. N. Jensen and A. Pandey, Trends Biotechnol., 2002, 20, 261–268.
8. F. J. Wang, C. X. Song, K. Cheng, X. N. Jiang, M. L. Ye and H. F. Zou, Anal. Chem., 2011, 83, 8078–8085.
9. B. Salovska, A. Tichy, M. Rezacova, J. Vavrova and E. Novotna, Rev. Anal. Chem., 2012, 31, 29–41.
10. X. S. Li, G. T. Zhu, Y. B. Luo, B. F. Yuan and Y. Q. Feng, TrAC, Trends Anal. Chem., 2013, 45, 233–247.
11. T. E. Thingholm, O. N. Jensen and M. R. Larsen, Proteomics, 2009, 9, 1451–1468.
12. J. D. Dunn, G. E. Reid and M. L. Bruening, Mass Spectrom. Rev., 2010, 29, 29–54.
13. I. L. Batalha, C. R. Lowe and A. C. A. Roque, Trends Biotechnol., 2012, 30, 100–110.
14. X. S. Li, L. D. Xu, G. T. Zhu, B. F. Yuan and Y. Q. Feng, Analyst, 2012, 137, 959–967.
15. A. Leitner, TrAC, Trends Anal. Chem., 2010, 29, 177–185.
16. L. Han, Z. Shan, D. H. Chen, X. J. Yu, P. Y. Yang, B. Tu and D. Y. Zhao, J. Colloid Interface Sci., 2008, 318, 315–321.
17. A. Lee, H. J. Yang, E. S. Lim, J. Kim and Y. Kim, Rapid Commun. Mass Spectrom., 2008, 22, 2561–2564.
18. M. Sturm, A. Leitner, J. H. Smatt, M. Linden and W. Lindner, Adv. Funct. Mater., 2008, 18, 2381–2389.
19. H. Y. Lin, W. Y. Chen and Y. C. Chen, Anal. Bioanal. Chem., 2009, 394, 2129–2136.
20. N. Hasan and H. F. Wu, Anal. Bioanal. Chem., 2011, 400, 3451–3462.
21. H. K. Kweon and K. Hakansson, Anal. Chem., 2006, 78, 1743–1749.
22. K. Imami, N. Sugiyama, Y. Kyono, M. Tomita and Y. Ishirama, Anal. Sci., 2008, 24, 161–166.
23. A. Leitner, M. Sturm, O. Hudecz, M. Mazanek, J. H. Smatt, M. Linden, W. Lindner and K. Mechtler, Anal. Chem., 2010, 82, 2726–2733.
24. J. N. M. Ballard, G. A. Lajoie and K. K. C. Yeung, J. Chromatogr. A, 2007, 1156, 101–110.
25. A. Schmidt, E. Csaszar, G. Ammerer and K. Mechtler, Proteomics, 2008, 8, 4577–4592.
26. C. K. Chang, C. C. Wu, Y. S. Wang and H. C. Chang, Anal. Chem., 2008, 80, 3791–3797.
27. T. E. Thingholm, O. N. Jensen, P. J. Robinson and M. R. Larsen, Mol. Cell. Proteomics, 2008, 7, 661–671.
28. L. A. Qiao, H. Y. Bi, J. M. Busnel, M. Hojeij, M. Mendez, B. H. Liu and H. H. Girault, Chem. Sci., 2010, 1, 374–382.
29. X. L. Li, Z. M. Guo, Q. Y. Sheng, X. Y. Xue and X. M. Liang, Analyst, 2012, 137, 2774–2776.
30. L. P. Li, T. Zheng, L. N. Xu, Z. Li, L. D. Sun, Z. X. Nie, Y. Bai and H. W. Liu, Chem. Commun., 2013, 49, 1762–1764.
31. H. J. Fu, Y. Q. Feng, Q. H. Zhang, S. L. Da and Y. J. Zhang, Anal. Lett., 1999, 32, 2761–2774.
32. Y. Q. Feng, H. J. Fu, Q. H. Zhang, S. L. Da and Y. J. Zhang, Chromatographia, 2000, 52, 165–168.
33. Y. Q. Feng, Q. H. Zhang, S. L. Da and Y. Zhang, Anal. Sci., 2000, 16, 579–583.
34. T. Li, Q. W. Yu, B. Lin and Y. Q. Feng, J. Sep. Sci., 2007, 30, 804–812.
35. H. B. He, Q. W. Yu, Y. Q. Feng and S. L. Da, J. Liq. Chromatogr. Relat. Technol., 2009, 32, 468–482.
36. V. Bolis, C. Morterra, M. Volante, L. Orio and B. Fubini, Langmuir, 1990, 6, 695–701.
37. C. Morterra and L. Orio, Mater. Chem. Phys., 1990, 24, 247–268.
38. A. Gervasini and A. Auroux, J. Catal., 1991, 131, 190–198.
39. J. Nawrocki, M. P. Rigney, A. McCormick and P. W. Carr, J. Chromatogr. A, 1993, 657, 229–282.
40. K. I. Hadjiivanov and D. G. Klissurski, Chem. Soc. Rev., 1996, 25, 61–69.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra13878c

---