Letter pubs.acs.org/ac
Surface Enhanced Laser Desorption Ionization of Phospholipids on Gold Nanoparticles for Mass Spectrometric Immunoassay Xiang-Cheng Lin, Xiang-Nan Wang, Lan Liu, Qian Wen,* Ru-Qin Yu, and Jian-Hui Jiang* State Key Laboratory of Chemeo/Bio-Sensing and Chemometrics, Institute of Chemical Biology and Nanomedicine, College of Chemistry and Chemical Engineering, Hunan University, Changsha, Hunan 410082, P. R. China S Supporting Information *
ABSTRACT: High-throughput and sensitive detection of proteins are essential for clinical diagnostics and biomarker discovery. We develop a novel high-throughput, multiplexed, sensitive mass spectrometric (MS) immunoassay method, which utilizes antibody-modified phospholipid bilayer coated gold nanoparticles (PBL-AuNPs) as the detection label and antibody-immobilized magnetic beads as the capture reagent. This method enables magnetic enrichment of the PBL-AuNPs label specific to target protein, allowing sensitive surface enhanced laser desorption ionization (SELDI)-TOF MS detection of the protein via its specific label. AuNPs act as not only the support but also the matrix for the phospholipids in SELDI TOF MS detection. Moreover, with phospholipids with varying molecular weights as the encoded MS reporters, this method allows multiplexed detection of multiple proteins. With the use of a predefined phospholipids internal standard, this method also affords excellent reproducibility in protein quantification. We have demonstrated this method using the assays of two tumor biomarkers, and the results reveal that it provides a sensitive platform for multiplexed protein detection with detection limits in the picomolar ranges. This method may provide a useful platform for high-throughput and sensitive detection of protein biomarkers for clinical diagnostics.
H
accurate atomic weight or molecular structural information.10,11 These advantages have led to increasing applications in proteomics and metabolomics.12,13 In recent years, we have also witnessed the growing interest in protein assays with metal element labels by using inductively coupled plasma mass spectrometry (ICPMS) detection14 and secondary ion MS imaging.15 The motivation for metal element based protein assays lies in the intrinsic high multiplexity capability in MS detection. However, the use of organic chemical labels for MS based protein detection has rarely been explored. There is a well-established MS based protein detection technology,16 which only utilizes the antibodies or aptamers to capture target proteins and uses MALDI-TOF MS for directly detecting the protein or its peptide fragments from enzymatic digestion. This technology has the advantages of improved selectivity and increased throughput, because MALDI-TOF MS allows the confirmation of the identity of target protein based on its characteristic fragments in very high throughput. Nevertheless, the detection of protein or its fragments may exhibit inferior sensitivity, which limits the applications of the MS based protein detection technology in assays for rare protein biomarkers.
igh-throughput and sensitive detection of proteins are essential for clinical diagnostics and biomarker discovery. Current methods for protein detection typically rely on specific affinity reagents such as antibodies and aptamers to discriminate target protein from coexisting components.1−3 The molecular recognition events between the affinity reagent and target protein are then detected using signal transduction technologies. Most of the detection technologies require the use of labels carrying optical, electrical, or electrochemical signals.4−6 A major effort in the assay design for protein detection is toward the development of high-throughput platforms that enable simultaneous detection of multiple proteins with enhanced sensitivity. This effort has catalyzed the evolution of different multiplex optical labels such as multicolor quantum dots (QDs),7 surface enhanced Raman scattering (SERS) tags,8 and upconversion luminescence (UCL) nanoparticles.9 However, the multiplexity of these optical labels are still limited because of their band-like spectrum. Moreover, the synthesis of these multiplex optical labels usually requires complicated steps with stringent control of the reaction conditions. Development of highly multiplexed, easily synthesized labels remains a critical pursuit in analytical chemistry and biomedicine. Mass spectrometry (MS) provides a powerful platform for detecting metal elements, small molecular compounds, and macromolecules such as nucleic acid and proteins, because of its super sensitivity, high throughput, and ability to offer © 2016 American Chemical Society
Received: July 17, 2016 Accepted: September 19, 2016 Published: September 19, 2016 9881
DOI: 10.1021/acs.analchem.6b02733 Anal. Chem. 2016, 88, 9881−9884
Letter
Analytical Chemistry
label, the immunoassay method is realized through a magnetic capture based sandwiched immunoassay (Scheme 1B). The capture antibody is tethered on the magnetic beads using an oriented immobilization chemistry via high affinity interaction between the antibody and protein G.22 Such an oriented immobilization of antibodies has been reported to substantially improve their target binding when compared to random orientations.22 In the presence of target protein, PBL-AuNPs label can be captured by the antibody-immobilized magnetic beads via a sandwiched complex between the capture antibody, target protein, and the detection antibody. After washing away unbound components in the sample and excessive labels, the magnetic beads are enriched and mixed with an internal standard followed by SELDI-TOF MS detection. The MS signals corresponding to the specific reporter phospholipids are then collected as indicators for quantification of the protein targets. To demonstrate the developed method, we choose two protein tumor biomarkers, prostate specific antigen (PSA) and carcino-embryonic antigen (CEA), as the case of study. To this end, we prepared two antibody-conjugated PBL-AuNPs labels, respectively, for these two biomarkers with their corresponding monoclonal antibodies using a two-step procedure. First, two PBL-AuNPs labels were synthesized readily using citratestabilized AuNPs (∼20 nm) via mixed self-assembly of two kinds of phospholipids, the click chemistry linker of 1,2distearoyl-sn-glycero-3-phosphoethanol-amine-N-[azido-polyethylene glycol-2000] (DSPE-PEG-Azide), and the MS reporters, 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) and 1,2-diarachidoyl-sn-glycero-3-phosphocholine (DAPC), respectively, for PSA and CEA. Then, the antibody-conjugated labels were synthesized via a copper-free strain-promoted azide−alkyne click chemistry reaction (SPAAC)23 by simply incubating dibenzocyclooctyne (DBCO)-modified monoclonal antibodies with the as-prepared PBL-AuNPs. The as-prepared antibody-conjuagted PBL-AuNPs labels were found to be dispersed very stably in 10 mM phosphate buffer (PB) containing 0.3 M NaCl and human serum (Figure S1), indicating that the PBL coating rendered high water dispersibility for the nanoparticles. Transmission electron microscope (TEM) measurements showed that the uncoated AuNPs had a diameter of ∼20 nm, while the PBL-AuNPs gave a very thin lipid coating of ∼4 nm (Figure 1A), a typical thickness for PBL.24 Dynamic light scattering (DLS) analysis indicated an increase in the average hydrodynamic diameters from the uncoated AuNPs (∼27 nm) to the PBL-AuNPs (∼48 nm) and the antibody-conjugated PBL-AuNPs (∼53 nm) (Figure S2). Zeta potential analysis evidenced a change in the surface charge of AuNPs from −25 to −38 mV after PBL selfassembly (Figure S3). Direct observations for the surface coating on the AuNPs were obtained using SELDI-TOF MS analysis (Figures 1B and S4). Obvious peaks were observed in the PBL-AuNPs labels for characteristic fragments from the azide moiety (MS peaks at 131.0 and 175.1 m/z for [(OCH2CH2)2N3 + H]+ and [(OCH2CH2)3N3 + H]+, respectively) and from DSPC (MS peak at 790.6 m/z) or DAPC (MS peak at 846.7 m/z). Further conjugation with antibodies produced three characteristic peaks for the SPAAC product at 319.1, 359.1, and 375.1 m/z. Taken together, these results verified the successful synthesis of the PBL-AuNPs labels. With the antibody-conjugated PBL-AuNPs label, the immunoassay method was realized for the detection of PSA
Previous studies have demonstrated that AuNPs can function as a selective matrix for surface enhanced laser desorption ionization (SELDI)-TOF MS detection of biomolecules with the ability to eliminate matrix ion interference and to improve sample homogeneity.17,18 However, the use of AuNPs as the SELDI MS label has not yet been reported for immunoassay. Because of the facile chemistry of synthesizing and modifying AuNPs with different organic compounds, AuNPs modified with specific MS reporters may provide a highly sensitive and multiplexed label for protein detection. Motivated by this hypothesis, we develop a novel high-throughput, multiplexed, and sensitive MS based immunoassay method, which for the first time uses phospholipid bilayer coated gold nanoparticles (PBL-AuNPs) as the label for SELDI-TOF MS detection, as illustrated in Scheme 1. Scheme 1. Illustration of SELDI-TOF MS Immunoassay Using PBL-AuNPs Labelsa
a
(A) Design of PBL-AuNPs labels and capture magnetic beads. (B) Procedure for immunoassay.
The choice of phospholipids as MS reporters lies in the fact that phospholipids can be modified on AuNPs as a noncovalent coating via a simple mixed self-assembly procedure,19 which enables efficient ionization without high energy to break the covalent bonds. On the basis of this choice, the SELDI-TOF MS label is designed to have an AuNP core and a mixed selfassembled bilayer coating of two kinds of phospholipids (Scheme 1A). One kind of the phospholipids is the reporter phospholipid, which is used as the encoded MS reporter specific to a target protein for generating a signal for quantifying the protein. The other kind of the phospholipids is a PEGylated phospholipid with azide functionality, in which the PEG moiety is used for reducing nonspecific adsorption of proteins20 and the azide group serves as a linker for antibody conjugation. With these designs, the resulting PBL-AuNPs not only yield a highly stable suspension in aqueous solution with resistance to nonspecific adsorption but also render a simple chemistry for conjugation with a detection antibody via a click chemistry reaction.21 For different antibodies against distinct target protein, the PBL-AuNPs label is synthesized in the same way except for the use of a specific reporter phospholipid to encode the target protein in order to achieve multiplexed detection. With the synthesized antibody-modified PBL-AuNPs 9882
DOI: 10.1021/acs.analchem.6b02733 Anal. Chem. 2016, 88, 9881−9884
Letter
Analytical Chemistry
Figure 1. TEM (A) and MS (B) characterization of nanoparticles: (1) citrate-stabilized AuNPs, (2) PBL-AuNPs, (3) antibody-conjugated PBL-AuNPs. a−e indicates MS Peaks at 131.0, 175.1, 319.1, 359.1, and 375.1 m/z, respectively.
Figure 2. SELDI-TOF MS signals in immunoassay of PSA and CEA. (A) Typical MS spectrum in assays of blank (black), 50 ng/mL PSA (red), 250 ng/mL CEA (blue), and 50 ng/mL PSA and 250 ng/mL CEA (brown). (B) MS peak intensities for DSPC at 790.6 m/z and DAPC at 846.7 m/z in assays of proteins or FBA.
and CEA. In this assay, a reagent mixture of magnetic beads immobilized with anti-PSA and anti-CEA antibodies, the DSPC-encoded PBL-AuNPs label conjugated with anti-PSA antibody, and DAPC-encoded PBL-AuNPs label conjugated with anti-CEA antibodies were incubated with the sample, followed by collection of the magnetic beads for SELDI-TOF MS detection. In an initial trial, we found that repetitive assays of 50 ng/mL PSA showed that the intensities of DSPC peak at 790.6 m/z varied substantially (RSD 30.8%), leading to inferior reproducibility in the assay (Figure S5A). To improve the reproducibility, we introduced an internal standard (IS) phospholipid of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) in the collected beads followed by SELDI-TOF MS analysis, and the peak intensity at 734.6 m/z corresponding to the IS parent ion was normalized to a constant (103). Using this normalization, repetitive assays of 50 ng/mL PSA were found to exhibit desired reproducibility (RSD 3.6%), implying the use of an IS based procedure offering the possibility of protein quantification (Figure S5B). In addition, the laser energy for SELDI experiment was optimized to maximize the MS signal (Figure S6). We found that the ratios of MS signals for DSPC to those for internal standard DPPC showed little variations with the changed laser energy, but the largest MS signal for DSPC was obtained with laser energy of 3900 au. Using the IS based procedure, we then explored the developed immunoassay for multiplexed detection of PSA and CEA (Figure 2A). When the reagent mixture was incubated with a blank sample with 1 mg/mL bovine serum albumin (BSA), very weak background was obtained in the MS spectrum with no appreciable peak for DSPC and DAPC except the IS peak. In contrast, in the presence of 50 ng/mL PSA, an intense peak appeared at 790.6 m/z corresponding to the parent ion [M + H]+ of DSPC. These results suggested that the observation of the DSPC peak gave an indicator for the presence of PSA in the sample. On the other hand, in the assay of 250 ng/mL CEA, we observed a high peak corresponding to the parent ion [M + H]+ of DAPC at 846.7 m/z, implying that the peak intensity from DAPC enabled detection of CEA. Furthermore, in the assay of a sample containing 250 ng/mL CEA and 50 ng/mL PSA, we obtained two intense peaks for DSPC and DAPC, and the intensities of these two peaks showed little deviation from those obtained in separate assays
of PSA or CEA. This finding revealed the ability of our method for multiplexed detection of PSA and CEA. Additionally, it was observed that the MS signals were obtained exclusively for DSPC and DAPC, respectively, in separate assays of PSA and CEA. These results revealed that no scrambled signals arose from incubation of the reagent mixture, suggesting no appreciable lipid exchange between the PBL-AuNPs labels. Such negligible lipid exchange might be ascribed to the short incubation time (1 h) for the immunoassay, which minimized the interference from the very low-rate process of lipid exchange. To test the selectivity of the method, control experiments with the addition of other proteins such as human serum albumin (HSA), immunoglobulin G (Ig G), α-feto protein (AFP), transferrin (TF), and a complicated biological matrix fetal bovine serum (FBA) were performed (Figure 2B). In the cases, we also only obtained a very small background, implying the high specificity of our method for PSA and CEA detection. Moreover, assays for two mixtures with FSA and PSA (50 ng/ mL) or CEA (250 ng/mL) gave a MS signal without significant deviation from that obtained for the assays of 50 ng/mL PSA or 250 ng/mL CEA. This finding testifies that our method enabled selective detection of PSA in complicated matrices. Next, the ability of the developed immunoassay for quantitative analysis of PSA and CEA was investigated by incubation of the reagent mixture with the target protein of varying concentrations. With increasing concentrations of PSA in the samples, we found that the intensities of MS peak for DSPC increased gradually up to a PSA concentration of 50 ng/ mL (Figure 3A). A linear correlation was obtained for MS peak intensities for DSPC to PSA concentration in the range from 0.1 to 50 ng/mL with an estimated detection limit of 0.03 ng/ mL according to the 3σ rule (Figure S7 and Table S1). Likewise, the intensities of MS peak intensities for DAPC were also observed to increase linearly with CEA concentrations in the range from 0.5 to 250 ng/mL with a detection limit estimated to be 0.2 ng/mL (Figures 3B and S8 and Table S2). Such detection limits were much (∼100-fold) better than exisiting MS based protein detection technologies,17 implying 9883
DOI: 10.1021/acs.analchem.6b02733 Anal. Chem. 2016, 88, 9881−9884
Letter
Analytical Chemistry
■
PBL-AuNPs and capture antibody immobilized magnetic beads, and SELDI-TOF MS immunoassay, as well as additional figures (PDF)
AUTHOR INFORMATION
Corresponding Authors
*Fax: +86-731-88821916. E-mail:
[email protected]. *Fax: +86-731-88821916. E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS This work was supported by NSFC (21527810, 21190041, 21521063, 21505040).
Figure 3. SELDI-TOF MS spectral signals in assays of PSA (A) and CEA (B) of varying concentrations.
that our method provided a feasible platform for sensitive and multiplexed quantitative detection of proteins. To demonstrate the potential of the developed method for clinical sample assays, we then applied the method for detection of PSA and CEA in 20 serum samples and compared it to ELISA. The results showed that the concentrations of these two proteins determined using our method were almost consistent with those determined by ELISA, and the RSDs ranged from −9.6% to 11.3% (Figure S9 and Tables S3 and S4). These data revealed the potential for the developed method for PSA and CEA detection in complicated samples. In conclusion, we developed a novel multiplexed, sensitive MS based immunoassay method. This method was realized through a sandwiched immunoassay using antibody-modified magnetic beads as the capture reagent, while using PBL-AuNPs as the label for SELDI-TOF MS detection. With phospholipids of varying molecular weights, this method allowed multiplexed detection of multiple proteins. Because there was a large bulk of commercialized phospholipids with varying molecular weights, this method was able to afford high multiplexity in protein detection. This method was demonstrated for the detection of PSA and CEA, and the results showed that it provided a useful platform for reproducible, sensitive, and multiplexed detection of these proteins. This method was also successfully applied to the detection of PSA and CEA in 20 serum samples. Considering the intrinsic advantage of high throughput in SELDI-TOF MS analysis and magnetic beads based separation, this method should hold great potential for high-throughput and sensitive detection of protein biomarkers for clinical diagnostics.
■
REFERENCES
(1) de la Rica, R.; Stevens, M. M. Nat. Nanotechnol. 2012, 7, 821− 824. (2) Potyrailo, R. A.; Murray, A. J.; Nagraj, N.; Pris, A. D.; Ashe, J. M.; Todorovic, M. Angew. Chem., Int. Ed. 2015, 54, 2174−2178. (3) Li, F.; Zhang, H. Q.; Wang, Z. X.; Newbigging, A. M.; Reid, M. S.; Li, X. F.; Le, X. C. Anal. Chem. 2015, 87, 274−292. (4) Nie, H. G.; Liu, S. J.; Yu, R. Q.; Jiang, J. H. Angew. Chem., Int. Ed. 2009, 48, 9862−9866. (5) Zhang, K.; Zheng, D.; Hao, L. L.; Cutler, J. I.; Auyeung, E.; Mirkin, C. A. Angew. Chem., Int. Ed. 2012, 51, 1169−1172. (6) Zhang, H.; Li, F.; Dever, B.; Li, X. F.; Le, X. C. Chem. Rev. 2013, 113, 2812−1841. (7) Zrazhevskiy, P.; Gao, X. H. Nat. Commun. 2013, 4, 1619. (8) Wang, Y.; Tang, L. J.; Jiang, J. H. Anal. Chem. 2013, 85, 9213− 9220. (9) Kale, V.; Pakkila, H.; Vainio, J.; Ahomaa, A.; Sirkka, N.; Lyytikainen, A.; Talha, S. M.; Kutsaya, A.; Waris, M.; Julkunen, I.; Soukka, T. Anal. Chem. 2016, 88, 4470−4477. (10) Konermann, L.; Vahidi, S.; Sowole, M. A. Anal. Chem. 2014, 86, 213−232. (11) Liu, S.; Wang, Y. S. Chem. Soc. Rev. 2015, 44, 7829−7854. (12) Zhang, Y. Y.; Fonslow, B. R.; Shan, B.; Baek, M. C.; Yates, J. R. Chem. Rev. 2013, 113, 2343−2394. (13) Huan, T.; Li, L. Anal. Chem. 2015, 87, 7011−7016. (14) Liu, R.; Zhang, S. X.; Wei, C.; Xing, Z.; Zhang, S. C.; Zhang, X. R. Acc. Chem. Res. 2016, 49, 775−783. (15) Angelo, M.; Bendall, S. C.; Finck, R.; Hale, M. B.; Hitzman, C.; Borowsky, A. D.; Levenson, R. M.; Lowe, J. B.; Liu, S. D.; Zhao, S. C.; Natkunam, Y.; Nolan, G. P. Nat. Med. 2014, 20, 436−442. (16) Zhang, X. Y.; Zhu, S. C.; Xiong, Y.; Deng, C. H.; Zhang, X. M. Angew. Chem., Int. Ed. 2013, 52, 6055−6058. (17) Kuo, H. Y.; DeLuca, T. A.; Miller, W. M.; Mrksich, M. Anal. Chem. 2013, 85, 10635−10642. (18) Yan, B.; Kim, S. T.; Kim, C. S.; Saha, K.; Moyano, D. F.; Xing, Y. Q.; Jiang, Y.; Roberts, A. L.; Alfonso, F. S.; Rotello, V. M.; Vachet, R. W. J. Am. Chem. Soc. 2013, 135, 12564−12567. (19) Suga, K.; Yoshida, T.; Ishii, H.; Okamoto, Y.; Nagao, D.; Konno, M.; Umakoshi, H. Anal. Chem. 2015, 87, 4772−4780. (20) Walkey, C. D.; Olsen, J. B.; Guo, H. B.; Emili, A.; Chan, W. C. W. J. Am. Chem. Soc. 2012, 134, 2139−2147. (21) Thirumurugan, P.; Matosiuk, D.; Jozwiak, K. Chem. Rev. 2013, 113, 4905−4979. (22) Borlido, L.; Azevedo, A. M.; Roque, A. C. A.; Aires-Barros, M. R. Biotechnol. Adv. 2013, 31, 1374−1385. (23) Zhao, M. X.; Liu, Y. R.; Hsieh, R. S.; Wang, N.; Tai, W. Y.; Joo, K.; Wang, P.; Gu, Z.; Tang, Y. J. Am. Chem. Soc. 2014, 136, 15319− 15325. (24) Liu, S. J.; Wen, Q.; Tang, L. J.; Jiang, J. H. Anal. Chem. 2012, 84, 5944−5950.
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.6b02733. Experimental methods including materials and reagents, preparation of PBL-AuNPs, detection of antibody labeled 9884
DOI: 10.1021/acs.analchem.6b02733 Anal. Chem. 2016, 88, 9881−9884
