Fabrication of a novel magnetic yolk–shell Fe₃O₄@mTiO₂@mSiO₂ nanocomposite for selective enrichment of endogenous phosphopeptides from a complex sample

Abstract

A yolk–shell nanocomposite composed of a magnetic mesoporous anatase TiO₂ (Fe₃O₄@mTiO₂) core, a medium cavity and an outermost mesoporous silica (mSiO₂) shell was successfully fabricated. Due to a combination of the strong magnetic response, improved diffusion of peptides, numerous affinity sites towards phosphopeptides and the size-exclusion effect, the nanocomposite demonstrated high enrichment efficacy and selectivity towards endogenous phosphopeptides from human serum.

---

Phosphorylation is one of the most important and ubiquitous post-translational modifications. The phosphorylation of endogenous peptides is involved in lots of biological and pathological variations, which makes endogenous phosphopeptides potential biomarkers with high clinical sensitivity and specificity. However, due to the interference and cover of signal by high-abundance proteins and endogenous non-phosphopeptides in the complex sample, the mass spectrometry (MS) detection of endogenous phosphopeptides still remains challenging.¹

Selective enrichment of endogenous phosphopeptides exists as an effective way to exclude the interference and improve the precision and sensitivity of MS detection. Compared with other enrichment methods, metal oxide affinity chromatography (MOAC) has been reported to demonstrate higher binding selectivity and sensitivity to phosphate group. Among these metal oxides, anatase TiO₂ has been most widely utilized for selective enrichment of phosphopeptides from various enzymatic digests due to the higher affinity and more tolerance of many biochemical reagents.² However, because of the occupation and cover of affinity sites on anatase TiO₂ by some large-molecule phosphoproteins and non-phosphoproteins in the complex sample, the enrichment selectivity and efficacy towards endogenous phosphopeptides would be highly affected.³ Appearing as a molecular sieve, mesoporous SiO₂ has been applied in different systems to differentiate small and large molecules based on the size-exclusion effect.⁴ But due to the low-efficiency and time-consuming centrifugation separation, the practical application of mesoporous SiO₂ in the exclusion of proteins beneficial for efficient and selective enrichment of endogenous phosphopeptides is limited. Recently, the magnetic yolk–shell structure has been attracting more and more attentions. On one hand, the magnetic response would simplify the process of sample separation. On the other hand, the cavity would be beneficial for the diffusion of guest molecules and also enhance the contact area between the substrate and guest molecules.⁵ Attributed to all above unique properties, an integration of mesoporous SiO₂, anatase TiO₂ and a magnetic yolk–shell structure should better fulfil the requirement for selective enrichment of endogenous phosphopeptides from the complex sample.

Herein, for highly efficient and selective enrichment of endogenous phosphopeptides from the complex sample, we elaborately fabricated a novel magnetic yolk–shell nanocomposite (Scheme 1). Starting from Fe₃O₄ nanoparticles, we directly coated a layer of amorphous TiO₂, followed by solvothermal reaction assisted pore formation and crystallization of amorphous TiO₂ into anatase type (Fe₃O₄@mTiO₂). Subsequently, with the help of PVP the Fe₃O₄@mTiO₂ was encapsulated by a dense SiO₂ shell (Fe₃O₄@mTiO₂@nSiO₂), onto which another outmost mesoprosous SiO₂ shell was assembled by using CTAB as a structure director (Fe₃O₄@mTiO₂@nSiO₂@mSiO₂–CTAB). After the selective etching of the dense SiO₂ shell by Na₂CO₃ solution and calcination of CTAB and PVP, the final nanocomposite (YS Fe₃O₄@mTiO₂@mSiO₂) was obtained. The YS Fe₃O₄@mTiO₂@mSiO₂ was suitable for selective enrichment of endogenous phosphopeptides from the complex sample: (1) the outmost mSiO₂ shell would act as a size-dependent sieve to exclude large-molecule proteins out; (2) the medium cavity would improve the diffusion of endogenous phosphopeptides and enhance the contact area between endogenous phosphopeptides and the Fe₃O₄@mTiO₂ core; (3) the Fe₃O₄@mTiO₂ core with high surface area would not only provide numerous affinity sites to capture endogenous phosphopeptides but also simplify the process of sample separation.

Scheme 1 The schematic illustration of the fabrication process of YS Fe₃O₄@mTiO₂@mSiO₂.

Shown in Fig. 1, the diameter of as-synthesized Fe₃O₄@mTiO₂ core distributed around 200–300 nm (Fig. 1c). After encapsulation with the dense SiO₂ shell, a well-defined core–shell structure was obtained (Fig. 1d). Under the direction of CTAB, another outmost mSiO₂ shell was successfully assembled and the corresponding size increased to 500–600 nm (Fig. 1e). The yolk–shell structure appeared after the selective etching process, which was based on the unexpected differentiated dissolution behaviours, stemming from stabilization of the outmost mSiO₂ shell by CTAB (Fig. 1f).⁶ Through the experiment, it was found out the mass ratio of the Fe₃O₄@mTiO₂@nSiO₂@mSiO₂–CTAB and Na₂CO₃ was an indispensable factor for the successful fabrication of the final yolk–shell structure. Fixing the amount of Na₂CO₃ at 212 mg, we varied the added amount of Fe₃O₄@mTiO₂@nSiO₂@mSiO₂–CTAB. Based on the TEM results (Fig. 2), if the added amount of Fe₃O₄@mTiO₂@nSiO₂@mSiO₂–CTAB was limited, an etching of both dense SiO₂ and outmost mSiO₂–CTAB shells happened. On the contrast, when the added amount of Fe₃O₄@mTiO₂@nSiO₂@mSiO₂–CTAB was too much, the Na₂CO₃ solution was not strong enough to completely etch the dense SiO₂ shell away, leaving a small gap. So only when the added amount of Fe₃O₄@mTiO₂@nSiO₂@mSiO₂–CTAB was appropriate, can a well-defined yolk–shell structure be successfully developed. According to the XRD analysis, all diffraction peaks of original magnetic nanoparticles can be perfectly indexed to the magnetite phase of Fe₃O₄ (JCPDS 19-629) and the final nanoparticles showed additional diffraction peaks of anatase TiO₂ (JCPDS 21-1272) (Fig. S1, ESI†). The EDS element mapping illustrated the exact distribution of fundamental compositions, further identifying the successful fabrication of YS Fe₃O₄@mTiO₂@mSiO₂ (Fig. 3).

Fig. 1 The SEM image of (a) YS Fe₃O₄@mTiO₂@mSiO₂; the TEM images of (b) Fe₃O₄, (c) Fe₃O₄@mTiO₂, (d) Fe₃O₄@mTiO₂@nSiO₂, (e) Fe₃O₄@mTiO₂@nSiO₂@mSiO₂–CTAB and (f) YS Fe₃O₄@mTiO₂@mSiO₂.

Fig. 2 TEM images of nanoparticles after the selective etching process by Na₂CO₃ solution when the added amount of Fe₃O₄@mTiO₂@nSiO₂@mSiO₂–CTAB was (a) 50 mg, (b) 100 mg, (c) 150 mg and (d) 250 mg.

Fig. 3 The EDS element mapping of YS Fe₃O₄@mTiO₂@mSiO₂.

Then, the N₂ adsorption experiment was used to probe the detailed structure information about YS Fe₃O₄@mTiO₂@mSiO₂. The adsorption–desorption isotherms exhibited the typical IV pattern of the mesoporous material with a distinct H3 hysteresis loop, proving the ink-bottle-type pores in which large cavities are connected by narrow windows.⁷ The pore distribution showed three main pore diameters of 2.7, 3.9 and 9.1 nm, which could be ascribed to pores distributing within the outmost mSiO₂ shell and inner mesoporous anatase TiO₂, respectively (Fig. S2, ESI†). Besides, the Fe₃O₄@mTiO₂ was also conducted through the BET and BJH analyses. The pores resulting from the solvothermal reaction centralized at 3.7 and 9.5 nm, matching the above result well and the surface area reached 169.79 g m⁻², providing numerous affinity sites to capture endogenous phosphopeptides. The original Fe₃O₄ exhibited a saturation magnetization value of 58.51 emu g⁻¹. While after coating with another three shells, the value decreased to 10.76 emu g⁻¹. Attributed to subsequent processes of selective etching and calcination, the saturation magnetization value increased to 15.7 emu g⁻¹, which would favour the easy separation of YS Fe₃O₄@mTiO₂@mSiO₂ (Fig. S3, ESI†).

The YS Fe₃O₄@mTiO₂@mSiO₂ was supposed to efficiently and selectively enrich endogenous phosphopeptides from the complex sample (Scheme 2). At first, the enrichment effect was evaluated with simple samples. Without enrichment the tryptic digests of β-casein showed lots of high-density non-phosphopeptides and the signal of phosphopeptides was suppressed. However, after enrichment with YS Fe₃O₄@mTiO₂@mSiO₂, three significant peaks corresponding to phosphopeptides were identified. Compared with the commercial TiO₂ (surface area: 61.25 g m⁻², Fig. S8, ESI†), the YS Fe₃O₄@mTiO₂@mSiO₂ demonstrated the higher enrichment efficacy and selectivity (Fig. 4a–c), which may be a result of the higher surface area and the special yolk–shell structure. Inspired by this, a detection limit experiment was subsequently conducted. The intensity strength of enriched phosphopeptides was positively related to the primitive concentration of tryptic digests of β-casein. And the detection limit could reach as low as 3 fmol, confirming the high enrichment efficacy (Fig. 4c–e). To further identify the high selectivity of YS Fe₃O₄@mTiO₂@mSiO₂, different amounts of tryptic digests of BSA as the interference were spiked into the determined amount of tryptic digests of β-casein. The direct analysis of the mixture showed no phosphopeptides but numerous non-phosphopeptides. Brightly, after enrichment with YS Fe₃O₄@mTiO₂@mSiO₂, only phosphopeptides existed. Even if the molar ratio was 1 : 1000 (β-casein : BSA), phosphopeptides could still be selectively captured (Fig. S4, ESI†). For YS Fe₃O₄@mTiO₂@mSiO₂, the outmost mSiO₂ shell would act as a size-dependent sieve to exclude the large-molecule proteins out beneficial for efficient and selective enrichment of phosphopeptides. To test this functionality, we applied YS Fe₃O₄@mTiO₂@mSiO₂ and Fe₃O₄@mTiO₂ in parallel for the enrichment of a mixture of α-casein (model phosphoprotein) and tryptic digests of β-casein (a mimic of the complex sample, mass ratio: 100 : 1). After enrichment, the elution of YS Fe₃O₄@mTiO₂@mSiO₂ contained very little α-casein with respect to high-abundance α-casein in the elution of Fe₃O₄@mTiO₂. On the contrast, as for the case of small-molecule peptides, the situation was opposite (Fig. 5). In detail, although both YS Fe₃O₄@mTiO₂@mSiO₂ and Fe₃O₄@mTiO₂ could capture the phosphopeptides, the signal intensity of phosphopeptides in the elution of Fe₃O₄@mTiO₂ was much weaker. It was speculated α-casein would cover and occupy lots of affinity sites on the Fe₃O₄@mTiO₂ in view of the large amount, which mostly affected the enrichment efficacy of phosphopeptides. However, the diffusion of large-molecule α-casein (23 690 Da, radius of gyration is about 4.5 nm) through the narrow mesopore channels (2.7 nm) distributing within the outmost mSiO₂ shell on the YS Fe₃O₄@mTiO₂@mSiO₂ was relatively difficult, resulting in the size-exclusion effect to improve the enrichment efficacy.

Scheme 2 The process of selective enrichment of endogenous phosphopeptides from the complex sample with YS Fe₃O₄@mTiO₂@mSiO₂.

Fig. 4 The MALDI-TOF-MS analysis of tryptic digests of β-casein. (a) Direct analysis; (b) after enrichment with commercial TiO₂; after enrichment with YS Fe₃O₄@mTiO₂@mSiO₂ when the added amount was (c) 50 fmol, (d) 10 fmol and (e) 3 fmol. * indicates phosphopeptides and # indicates dephosphorylated peptides.

Fig. 5 The MALDI-TOF-MS analysis of the elution after enrichment of a mixture of α-casein protein and tryptic digests of β-casein with (a and c) Fe₃O₄@mTiO₂ and (b and d) YS Fe₃O₄@mTiO₂@mSiO₂.

After all, we applied the YS Fe₃O₄@mTiO₂@mSiO₂ in the selective enrichment of endogenous phosphopeptides from human serum. The direct analysis of human serum showed none of endogenous phosphopeptides, accompanied with lots of high-density endogenous non-phosphopeptides and proteins. However, after enrichment only four endogenous phosphopeptides with high signal-to-noise were detected, which showed better enrichment selectivity and efficacy than previous studies.⁸ In the high-molecular range, two obvious proteins (e.g. human serum albumin, 67 kDa, 5 × 7 × 7 nm) were found out in the supernatant with respect to nothing in the elution, all of which hinted the effective capacity of YS Fe₃O₄@mTiO₂@mSiO₂ for selective enrichment of endogenous phosphopeptides from the complex sample (Fig. 6).

Fig. 6 The MALDI-TOF-MS analysis of human serum. (a) Direct analysis; (b) the endogenous phosphopeptides in the elution, the residual proteins in the (c) elution and (d) supernatant after enrichment with YS Fe₃O₄@mTiO₂@mSiO₂. # indicates endogenous phosphopeptides and * indicates residual proteins in human serum.

In summary, a novel magnetic yolk–shell Fe₃O₄@mTiO₂@mSiO₂ nanocomposite was successfully developed. The nanocomposite exhibited a strong magnetic response, high enrichment selectivity and efficacy towards phosphopeptides and the size-exclusion effect, which was suitable for selective enrichment of endogenous phosphopeptides from the complex sample. With all these advances, it is believed the nanocomposite would be applied in many other fields in the future.

Acknowledgements

This work was supported by the China State Key Basic Research Program Grant (2013CB911202, 2012CB910604), the Creative Research Group Project of NSFC (21321064), the National Natural Science Foundation of China (21235006, 81161120540), National Key Special Program on Infection diseases (2012ZX10002009-011), Analytical Method Innovation Program of MOST (2012IM030900) to H. F. Zou.

Notes and references

1. (a) G. Cheng, Y. Wang, Z. Wang, X. Sui, J. Zhang and J. Ni, RSC Adv., 2014, 4, 7694–7702; (b) L. Zhao, H. Qin, Z. Hu, Y. Zhang, R. a. Wu and H. Zou, Chem. Sci., 2012, 3, 2828; (c) Y. Yan, Z. Zheng, C. Deng, Y. Li, X. Zhang and P. Yang, Anal. Chem., 2013, 85, 8483–8487; (d) J. C. Smith, M. A. Duchesne, P. Tozzi, M. Ethier and D. Figeys, J. Proteome Res., 2007, 6, 3174–3186.
2. (a) Y. Yan, X. Sun, C. Deng, Y. Li and X. Zhang, Anal. Chem., 2014, 86, 4327–4332; (b) Y. Liang, X. He, L. Chen and Y. Zhang, RSC Adv., 2014, 4, 18132–18135; (c) W. F. Ma, Y. Zhang, L. L. Li, L. J. You, P. Zhang, Y. T. Zhang, J. M. Li, M. Yu, J. Guo, H. J. Lu and C. C. Wang, ACS Nano, 2012, 6, 3179–3188; (d) T. E. Thingholm, T. J. Jorgensen, O. N. Jensen and M. R. Larsen, Nat. Protoc., 2006, 1, 1929–1935; (e) Y. Li, X. M. Zhang and C. H. Deng, Chem. Soc. Rev., 2013, 42, 8517–8539; (f) W. F. Ma, Y. T. Zhang, M. Yu, J. X. Wan and C. C. Wang, RSC Adv., 2014, 4, 9148–9151.
3. (a) Z. Lu, M. Ye, N. Li, W. Zhong and Y. Yin, Angew. Chem., Int. Ed., 2010, 49, 1862–1866; (b) G. Cheng, Z. G. Wang, Y. L. Liu, J. L. Zhang, D. H. Sun and J. Z. Ni, ACS Appl. Mater. Interfaces, 2013, 5, 3182–3190; (c) Y. Zhang, W. F. Ma, C. Zhang, C. C. Wang and H. J. Lu, ACS Appl. Mater. Interfaces, 2014, 6, 6290–6299.
4. (a) R. Tian, H. Zhang, M. Ye, X. Jiang, L. Hu, X. Li, X. Bao and H. Zou, Angew. Chem., Int. Ed., 2007, 46, 962–965; (b) A. Bouamrani, Y. Hu, E. Tasciotti, L. Li, C. Chiappini, X. Liu and M. Ferrari, Proteomics, 2010, 10, 496–505; (c) Y. Hu, A. Bouamrani, E. Tasciotti, L. Li, X. W. Liu and M. Ferrari, ACS Nano, 2010, 1, 439–451; (d) Z. Y. Gu, Y. J. Chen, J. Q. Jiang and X. P. Yan, Chem. Commun., 2011, 47, 4787–4789; (e) N. R. Sun, C. H. Deng, Y. Li and X. M. Zhang, ACS Appl. Mater. Interfaces, 2014, 6, 11799–11804.
5. (a) H. Wan, H. Qin, Z. Xiong, W. Zhang and H. Zou, Nanoscale, 2013, 5, 10936–10944; (b) W. Zhao, H. Chen, Y. Li, L. Li, M. Lang and J. Shi, Adv. Fun. Mater., 2008, 18, 2780–2788; (c) Y. Chen, H. R. Chen, L. M. Guo, Q. J. He, F. Chen, J. Zhou, J. W. Feng and J. L. Shi, ACS Nano, 2010, 1, 529–539.
6. (a) X. Fang, C. Chen, Z. Liu, P. Liu and N. Zheng, Nanoscale, 2011, 3, 1632–1639; (b) Y. Fang, G. Zheng, J. Yang, H. Tang, Y. Zhang, B. Kong, Y. Lv, C. Xu, A. M. Asiri, J. Zi, F. Zhang and D. Zhao, Angew. Chem., Int. Ed., 2014, 53, 5366–5370; (c) J. Yang, D. Shen, L. Zhou, W. Li, J. Fan, A. M. El-Toni, W. X. Zhang, F. Zhang and D. Zhao, Adv. Healthcare Mater., 2014 DOI:10.1002/adhm.201400053.
7. (a) Y. Yang, X. Liu, X. Li, J. Zhao, S. Bai, J. Liu and Q. Yang, Angew. Chem., Int. Ed., 2012, 51, 9164–9168; (b) H. Wan, Y. Zhang, Z. Y. Liu, G. J. Xu, G. Huang, Y. S. Ji, Z. C. Xiong, Q. Q. Zhang, J. Dong, W. B. Zhang and H. F. Zou, Nanoscale, 2014, 6, 8743–8753; (c) Y. Yang, J. Liu, X. Li, X. Liu and Q. Yang, Chem. Mater., 2011, 23, 3676–3684.
8. (a) G. Cheng, X. Yu, M. D. Zhou and S. Y. Zheng, J. Mater. Chem. B, 2014, 2, 4711–4719; (b) Z. Lu, J. C. Duan, L. He, Y. X. Hu and Y. D. Yin, Anal. Chem., 2010, 82, 7249–7258; (c) Z. G. Wang, G. Cheng, Y. L. Liu, J. L. Zhang, D. H. Sun and J. Z. Ni, J. Mater. Chem. B, 2013, 1, 1491–1500.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details and additional figures. See DOI: 10.1039/c4ra08692a

---