Letter pubs.acs.org/ac
Membrane Electrospray Ionization for Direct Ultrasensitive Biomarker Quantitation in Biofluids Using Mass Spectrometry Mei Zhang,†,‡,§ Fankai Lin,⊥ Jianguo Xu,†,‡,§ and Wei Xu*,⊥ †
National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, Beijing, China, 102206 ‡ State Key Laboratory for Infectious Disease Prevention and Control, Beijing, China, 102206 § Collaborative Innovation Center for Diagnosis and Treatment of Infectious Diseases, Hangzhou, China, 310003 ⊥ School of Life Science, Beijing Institute of Technology, Beijing, China, 100081 S Supporting Information *
ABSTRACT: The ability of rapid biomarker quantitation in raw biological samples would expand the application of mass spectrometry in clinical diagnosis. Up until now, the conventional chromatography−mass spectrometry method is time-consuming in both sample preparation and chromatography separation processes, while ambient ionization methods normally suffer from sensitivity. The membrane electrospray ionization (MESI) introduced in this study could not only achieve sensitive biomolecule quantitation, but also minimize the sample handling process. As a unique feature of MESI, both vertical and horizontal chemical separations could be achieved in real-time. With the capability of mass-selectively minimizing matrix effects from salts, small molecules, and macromolecules, ultrasensitive detection of cytochrome C (>500-fold sensitivity improvement) in raw urine samples was demonstrated in less than 20 min.
W
and organic matrices) could be mass-selectively decreased. With minimum sample pretreatment, quantitative analyses of organic compounds, peptides, and proteins were achieved for samples with salt concentrations at physiological levels. Using dialysis membranes coated with antibodies, fast targeted protein quantitation (cytochrome C) in raw urine samples was achieved at the subppb level, >500-fold sensitivity improvement compared with conventional nanoESI. In comparison, high sensitivity detections (ppb level) have been reported for PS, but mostly for small molecules, such as drugs.37−40 PS has not been used to ionize proteins until 2014, in which year, Y. Zhang et al. investigated the structures of noncovalent protein complexes ionized by PS.38,46 The vertical separation capability of MESI was also demonstrated for the detection of multiple target molecules using a dual-layer membrane setup.
ith high sensitivity and specificity, mass spectrometry (MS) has been widely used in chemical identifications and detections. When coupled with separation techniques (liquid chromatography, LC, for example), it allows qualitative and quantitative analyses of trace amount analytes, even within complex backgrounds.1−5 Recent developments of ambient ionization techniques, including desorption electrospray ionization (DESI),6 direct analysis in real-time (DART),7 paper spray ionization (PS),8 and many others,9−22 permit direct sample analyses with minimized sample preparation and/ or separation processes. Nevertheless, sensitivity and quantitation accuracy of both LC−MS and ambient ionization-MS techniques can be substantially affected by the matrix effects,23,24 especially for biological samples with both organic (e.g., sugars, amines, and pigments) and inorganic (e.g., salts) matrixes. Efforts have been made for online removal of salts in electrospray ionization (ESI) using methods such as online membrane cleanup,25−28 adding solution additives,29 and etc.30,31 A number of ambient ionization methods have also been developed for analysis of biomolecules within highly concentrated salt solutions.32−36 Following the idea of PS, this study presents membrane electrospray ionization (MESI), which could achieve biomolecule analysis in raw biofluids with high sensitivity. As a unique nature of MESI, both vertical and horizontal chemical separations could be achieved in real-time. Matrix effects from interferences not only at the high-mass end (macromolecules and cells, for instance) but also at the low-mass end (inorganic © XXXX American Chemical Society
■
EXPERIMENTAL SECTION All experiments were carried out using a Bruker HCT mass spectrometer (Bruker Daltonics Inc., MA, Germany). Cytochrome C, myoglobin, lysozyme, angiotensin II, peptide MetArg-Phe-Ala (MRFA), progesterone, and urea were purchased from Sigma-Aldrich (MO). Cytochrome C antibody (catalog no. MAB898) was purchased from R&D systems (MN). Inorganic compounds and filter paper were purchased from Received: November 25, 2014 Accepted: March 1, 2015
A
DOI: 10.1021/acs.analchem.5b00467 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry
Figure 1. (a) On-membrane matrix removal and analyte ionization in MESI. (b) Scanning electron microscopy photos of the membrane before and after MESI.
Figure 2. Mass spectra of 10 μg mL−1 MRFA in PBS (100 mM, pH 7.4) (a) using MESI and (b) using nanoESI. Mass spectra of 500 μg mL−1 cytochrome C in 0.85% NaCl (c) using MESI and (d) using nanoESI. SNR comparisons between MESI and nanoESI for (e) 10 μg mL−1 MRFA and (f) 500 μg mL−1 cytochrome C with respect to different salt concentrations. Linear detection ranges for (g) MRFA in 0.85% NaCl and (h) cytochrome C in 0.85% NaCl. Membranes with MWCO 1 kDa were used for MRFA and 3.5 kDa for cytochrome C.
kV) and flushing solvent (methanol/water 6:4 v/v, 0.1% formic acid, 8 μL) are applied onto the membrane. MESI operation condition optimization process could be found in the Supporting Information. Molecules remaining on the membrane are transported and ionized at the tip of the triangular membrane. Because of the limited amount of sample and solvent loaded on the membrane, ion current could last ∼1 min (Figure S2 in the Supporting Information). The mass spectrum at the peak tops of the extracted ion chromatogram (EIC, m/z value of the target molecule was extracted) of MESI were chosen and used in this work. Figure 1b shows the scanning electron microscopy images on the membrane before and after flushing, in which almost all samples (progesterone in saline) were depleted after being flushed.
Beijing Chemical Company (Beijing, China). Dialysis membranes were purchased from Ebioeasy Co., Ltd. (Shanghai, China). Human blood, tear, urine, and saliva were provided by healthy volunteers in accordance with the requirements of medical ethics. In a MESI experiment (shown in Figure 1a and Figure S1 in the Supporting Information), a 2 μL sample is first loaded on a piece of pretreated triangular dialysis membrane (∼7 mm equilateral triangle, and details about the pretreatment of the membrane could be found in the Supporting Information), which was tightly covered on a piece of prewetted filter paper. The tip of the membrane is placed about 5 mm away from the MS inlet, and there is a 10−20° angle between the membrane surface and the horizontal plane to facilitate the movement of liquid on the membrane. In about 30 s, a high dc voltage (∼3.5 B
DOI: 10.1021/acs.analchem.5b00467 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry
Figure 3. Direct analyses of (a) cytochrome C in human tears, (b) angiotensin II in urine, (c) MRFA in human serum, and (d) progesterone in human saliva. Membranes with MWCO 1 kDa were used for MRFA and progesterone and 3.5 kDa for cytochrome C and angiotensin II (see text for details).
■
Because of the charge competition effects in ESI sources,41 the MS sensitivity and signal-to-noise ratio (SNR) would be substantially affected by the presence of nonvolatile organic and inorganic salts. With the help of a dialysis membrane, MESI could effectively diminish salts and achieve both high sensitivities and high SNRs for the MS analyses of biomolecules. As an example, Figure 2 compares the performances of MESI and nanoESI when analyzing MRFA and cytochrome C in solutions with salts. Working conditions of the nanoESI used in this study were optimized independently in terms of applied voltage and distance, which were then chosen to be ∼1 kV and 7 mm, respectively after optimization. Presented in Figure 2a−d, MESI could effectively improve the SNRs of mass spectra by reducing the noises from salts, thus minimizing the charge competition effects. The SNR improvements at different salt concentrations and for different analytes were plotted in parts e and f of Figure 2. Generally, the SNR enhancement is more prominent for samples with higher salt concentrations and larger analytes. About 10−20-fold SNR enhancements were observed for cytochrome C in NaCl solutions (1−5%). As shown in Figure 2g,h, the limit of quantitation (LOQ) for MRFA and cytochrome C in 0.85% NaCl solution were 10 ng mL−1 and 500 ng mL−1, respectively, when using MESI. These LOQ values were 10-fold lower than those obtained from nanoESI. Results for other compounds, as well as direct comparisons with PS, could be found in the Supporting Information (Figures S9 and S10). These results indicate that MESI offers a wider linear range and lower LOQ, especially for peptides or proteins whose molecular masses were relatively large (compared with the molecular weights of salts). A possible reason is membrane with larger pore sizes could be applied for larger targeted molecules, which could enhance the removal of salts and small interference molecules in the samples (Figure S4 in the Supporting Information).
RESULTS AND DISCUSSION The ionization mechanism of MESI was similar to that of PS, which was assisted by the high electric field generated at the tip of the triangle. However, the sample separation process in MESI is different from that of PS. In MESI, both vertical and horizontal chemical separations would happen at the same time, and ions were separated vertically based on their physical sizes. After samples being loaded on the membrane, salts and molecules with sizes smaller than the pore size of the membrane (or molecular weight cutoff, MWCO) could diffuse through the membrane and be absorbed by the filter paper. Figure S3 in Supporting Information monitored the dynamic salt removing process in a MESI, suggesting salts could be effectively removed in ∼15 s after applying the flushing solvent. Conventionally, it takes hours to perform a dialysis experiment. In a MESI experiment, the filter paper, gravity, electric driven force, and membranes with large pore sizes are believed to help the diffusion of small molecules through membranes in a much shorter period. By selecting the dialysis membrane with different MWCOs, different molecules could be massselectively retained on top of the membrane, as shown in Figure S4 in the Supporting Information. Optimization of the operation procedures and selection of different membranes for different targeted molecules could be found in the Supporting Information (Figures S5−S8). With cellulose-based microstructures, PS would have better separation effects horizontally, especially for matrix/interference molecules with high masses,8 which could not be filtered by membranes in the vertical direction. Furthermore, since the membrane is hydrophobic, membrane with a much larger MWCO was used to help the spread and transportation of solutions. The prewetted filter paper attached to the membrane also helps in terms of connecting liquid on the membrane with the high voltage, especially after the liquid on the membrane moved forward due to the consumption of spray and evaporation. C
DOI: 10.1021/acs.analchem.5b00467 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry
Figure 4. Ultrasensitive cytochrome C detection in human urine in 20 min: (a) MESI methodology using membranes coated with antibody; (b) the corresponding linear detection range of cytochrome C; (c) mass spectrum of 10 ng mL−1 (LOQ) cytochrome C in urine detected by MESI coated with anticytochrome C; (d) mass spectrum of 500 μg mL−1 cytochrome C in urine detected by MESI without anticytochrome C.
Figure 5. MESI with dual membrane layers for multiple target biomolecule analyses. (a) Diagram of MESI with dual membrane. (b) Cytochrome C detected from the first layer. (c) MRFA detected from the second layer and (d) the corresponding tandem mass spectrum.
With remarkable tolerance to matrix effects, MESI could be applied to the direct analyses of biofluids with minimum or no sample preparations. The detection of cytochrome C in human tears, angiotensin II in urine, MRFA in serum (diluted 10 times with water), and progesterone in human saliva could be
achieved with results shown in Figure 3. Cytochrome C and angiotensin II powders were directly spiked into human tears and urine, respectively. MRFA powder was spiked into 10-time diluted serum with water. A 200 μg mL−1 progesterone solution was first prepared in methanol and then mixed with human D
DOI: 10.1021/acs.analchem.5b00467 Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
■
saliva with a 1:1 ratio (v/v). A recovery rates between 97.4− 104.2% and a RSD < 2% were achieved for the biomolecules tested above (Table S1 in Supporting Information). In addition, MESI could also work in the negative mode as shown in Figure S11 in the Supporting Information. The detection sensitivities of targeted biomolecules could be further improved by coating the microdialysis membrane with antibodies. It was reported that cytochrome C in human urine can be utilized as a clinical biomarker for the diagnosis of druginduced liver injury.42 As shown in Figure 4a, the microdialysis membrane (MWCO 3.5 kDa) was first coated with anticytochrome C and then dipped into human urine samples spiked with cytochrome C at 37 °C for 15 min to allow the binding of cytochrome C with anticytochrome C. After washing 3 times by pure water, 10 μL 7 mol L−1 urea was added onto the membrane to separate antibody and antigen. Ionized using MESI, a LOQ of 10 ng mL−1 (500-fold improvement compared with 50 μg mL−1 using nanoESI in Figure 2h) and a linear quantitation range from 10 to 1250 ng mL−1 could be achieved (Figure 4b), which covers the entire clinical concentration range.42,43 Potentially, MESI coupled with antibody could also be applied to high-throughput antibody screening and specificity evaluations.44,45 Figure 4c,d showed mass spectra of cytochrome C detected by MESI coated with and without anticytochrome C in human urine, respectively, which suggested that immunoadsorption on the membrane can provide significant sensitivity improvement. Besides monitoring (or optimized for) one biomolecule, multiple biomolecules could be targeted at the same time using multiple layers of membranes with different pore sizes and pointing at different directions (Figure 5a). As an example, a mixture of cytochrome C (1 mg mL−1) and MRFA (100 μg mL−1) in 0.85% NaCl was loaded on a dual layer MESI source. A membrane with 3.5 kDa MWCO was chosen as the first layer, which stops cytochrome C; while the second layer has a MWCO of 1 kDa, which would retain MRFA. By facing the first membrane toward a MS, cytochrome C could be detected after flushing solvent onto the first layer (Figure 5b). By removing the first layer and turning the second membrane toward the MS, MRFA could then be flushed and detected (Figure 5c,d).
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Phone: +86-10-68918123. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We are grateful for the suggestions from Prof. Graham R. Cooks of Purdue University. This research was supported by NNSF China (Grant 21205005, 81471919), MOST China (Grant 2011YQ0900502), 1000 plan, Research Foundation of China CDC (Grant 2014A101).
■
REFERENCES
(1) Boggio, K. J.; Obasuyi, E.; Sugino, K.; Nelson, S. B.; Agar, N. Y.; Agar, J. N. Expert Rev. Proteomics 2011, 8, 591−604. (2) Date, S.; Mizuno, H.; Tsuyama, N.; Harada, T.; Masujima, T. Anal. Sci. 2012, 28, 201−203. (3) Mellors, J. S.; Jorabchi, K.; Smith, L. M.; Ramsey, J. M. Anal. Chem. 2010, 82, 967−973. (4) Dimri, G. P.; Lee, X.; Basile, G.; Acosta, M.; Scott, G.; Roskelley, C.; Medrano, E. E.; Linskens, M.; Rubelj, I.; Pereira-Smith, O. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 9363−9367. (5) Nemes, P.; Knolhoff, A. M.; Rubakhin, S. S.; Sweedler, J. V. Anal. Chem. 2011, 83, 6810−6817. (6) Takáts, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306, 471−473. (7) Cody, R. B.; Laramée, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 2297−2302. (8) Wang, H.; Liu, J.; Cooks, R. G.; Ouyang, Z. Angew. Chem., Int. Ed. 2010, 49, 877−880. (9) Venter, A.; Nefliu, M.; Cooks, R. G. TrAC, Trends Anal. Chem. 2008, 27, 284−290. (10) Westphall, M. S.; Jorabchi, K.; Smith, L. M. Anal. Chem. 2008, 80, 5847−5953. (11) Chetwani, N.; Cassou, C. A.; Go, D. B.; Chang, H. Anal. Chem. 2011, 83, 3017−3023. (12) Na, N.; Zhao, M.; Zhang, S.; Yang, C.; Zhang, X. J. Am. Soc. Mass. Spectrom. 2007, 18, 1859−1862. (13) Wei, Z.; Han, S.; Gong, X.; Zhao, Y.; Yang, C.; Zhang, S.; Zhang, X. Angew. Chem., Int. Ed. 2013, 52, 11025−11028. (14) Gong, X.; Zhao, Y.; Cai, S.; Fu, S.; Yang, C.; Zhang, S.; Zhang, X. Anal. Chem. 2014, 86, 3809−3816. (15) Vaikkinen, A.; Shrestha, B.; Kauppila, T. J.; Vertes, A.; Kostiainen, R. Anal. Chem. 2012, 84, 1630−1636. (16) Meier, L.; Berchtold, C.; Schmid, S.; Zenobi, R. Anal. Chem. 2012, 84, 2076−2080. (17) Haapala, M.; Suominen, T.; Kostiainen, R. Anal. Chem. 2013, 85, 5715−5719. (18) Liu, P.; Forni, A.; Chen, H. Anal. Chem. 2014, 86, 4024−4032. (19) Gilbert-Lopez, B.; Schilling, M.; Ahlmann, N.; Michels, A.; Hayen, H.; Molina-Díaz, A.; García-Reyes, J. F.; Franzke, J. Anal. Chem. 2013, 85, 3174−3182. (20) Zhan, X.; Zhao, Z.; Yuan, X.; Wang, Q.; Li, D.; Xie, H.; Li, X.; Zhou, M.; Duan, Y. Anal. Chem. 2013, 85, 4512−4519. (21) Gong, M.; Zhang, P.; MacDonald, B. D.; Sinton, D. Anal. Chem. 2014, 86, 8090−8097. (22) Wang, B.; Trimpin, S. Anal. Chem. 2014, 86, 1000−1006. (23) Taylor, P. J. Clin. Biochem. 2005, 38, 328−334. (24) King, R.; Bonfiglio, R.; Fernandez-Metzler, C.; Miller-Stein, C.; Olah, T. J. Am. Soc. Mass Spectrom. 2000, 11, 942−950. (25) Tibavinsky, I. A.; Kottke, P. A.; Fedorov, A. G. Anal. Chem. 2015, 87, 351−356. (26) Chen, Y.; Mori, M.; Pastusek, A. C.; Schug, K. A.; Dasgupta, P. K. Anal. Chem. 2011, 83, 1015−1021. (27) Sun, L.; Duan, J.; Tao, D.; Liang, Z.; Zhang, W.; Zhang, L.; Zhang, Y. Rapid Commun. Mass Spectrom. 2008, 22, 2391−2397.
■
CONCLUSIONS In summary, membrane electrospray ionization was introduced as a novel ambient ionization technique, and MESI-MS has satisfactory performances for targeted biomarker quantitation in raw biofluids. This technique shows great sensitivity for various biomolecule analyses, especially for large biomolecules (peptides or proteins) in very complex matrixes. With the capability of quantifying trace amount biomarkers in raw biofluids in real-time, MESI could accelerate the translation of mass spectrometry techniques to clinical applications.
■
Letter
ASSOCIATED CONTENT
S Supporting Information *
Materials and methods, investigation of the total ion chromatogram, selection of dialysis membranes, optimization of experimental conditions, comparison of MESI with nanoESI and PS, methodology validation, negative mode, and MESI using membrane coated with antibody. This material is available free of charge via the Internet at http://pubs.acs.org. E
DOI: 10.1021/acs.analchem.5b00467 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry (28) Chung, Y. T.; Ling, Y. C.; Yang, C. S.; Sun, Y. C.; Lee, P. L.; Lin, C. Y.; Hong, C. C.; Yang, M. H. Anal. Chem. 2007, 79, 8900−8910. (29) Mridul, K. M.; Lee, C. C.; Yutaka, H.; Zhan, Y.; Kenzo, H. Anal. Methods 2010, 2, 1905−1912. (30) Cheng, C.; Chen, C. S.; Hsieh, P. H. J. Chromatogr. A 2010, 1217, 2104−2110. (31) Ji, X.; Li, D.; Li, H. Biomed. Chromatogr. 2015, DOI: 10.1002/ bmc.3418. (32) Chang, D. Y.; Lee, C. C.; Shiea, J. Anal. Chem. 2002, 74, 2465− 2469. (33) Han, F.; Yang, Y.; Ouyang, J.; Na, N. Analyst 2015, 140, 710− 715. (34) Jackson, A. U.; Talaty, N.; Cooks, R. G.; Van Berkel, G. J. J. Am. Soc. Mass Spectrom. 2007, 18, 2218−2225. (35) Chen, H.; Venter, A.; Cooks, R. G. Chem. Commun. (Cambridge, U. K.) 2006, 19, 2042−2044. (36) Chen, H.; Yang, S.; Li, M.; Hu, B.; Li, J.; Wang, J. Angew. Chem., Int. Ed. 2010, 49, 3053−3056. (37) Lin, C.; Liao, W.; Chen, H.; Kuo, T. Bioanalysis 2014, 6, 199− 208. (38) Zhang, Y.; Ju, Y.; Huang, C.; Wysocki, V. H. Anal. Chem. 2014, 86, 1342−1346. (39) Wang, H.; Manicke, N. E.; Yang, Q.; Zheng, L.; Shi, R.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2011, 83, 1197−1201. (40) Wang, H.; Ren, Y.; McLuckey, M. N.; Manicke, N. E.; Park, J.; Zheng, L.; Shi, R.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2013, 85, 11540−11544. (41) Tang, K.; Page, J. S.; Smith, R. D. J. Am. Soc. Mass Spectrom. 2004, 15, 1416−1423. (42) Miller, T. J.; Knapton, A.; Adeyemo, O.; Noory, L.; Weaver, J.; Hanig, J. P. J. Appl. Toxicol. 2008, 28, 815−828. (43) Small, D. M.; Gobe, G. C. Expert Opin. Drug Metab. Toxicol. 2012, 8, 655−664. (44) Janda, K. D.; Schloeder, D.; Benkovic, S. J.; Lerner, R. A. Science 1988, 241, 1188−1191. (45) Roberts, S.; Cheetham, J. C.; Rees, A. R. Nature 1987, 328, 731−734. (46) Liu, J.; Wang, H.; Manicke, N. E.; Lin, J. M.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2010, 82, 2463−2471.
F
DOI: 10.1021/acs.analchem.5b00467 Anal. Chem. XXXX, XXX, XXX−XXX
Membrane Electrospray Ionization for Direct Ultrasensitive Biomarker Quantitation in Biofluids Using Mass Spectrometry Mei Zhang,†,‡,§ Fankai Lin,⊥ Jianguo Xu,†,‡,§ and Wei Xu*,⊥ †
National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, Beijing, China, 102206 ‡ State Key Laboratory for Infectious Disease Prevention and Control, Beijing, China, 102206 § Collaborative Innovation Center for Diagnosis and Treatment of Infectious Diseases, Hangzhou, China, 310003 ⊥ School of Life Science, Beijing Institute of Technology, Beijing, China, 100081 S Supporting Information *
ABSTRACT: The ability of rapid biomarker quantitation in raw biological samples would expand the application of mass spectrometry in clinical diagnosis. Up until now, the conventional chromatography−mass spectrometry method is time-consuming in both sample preparation and chromatography separation processes, while ambient ionization methods normally suffer from sensitivity. The membrane electrospray ionization (MESI) introduced in this study could not only achieve sensitive biomolecule quantitation, but also minimize the sample handling process. As a unique feature of MESI, both vertical and horizontal chemical separations could be achieved in real-time. With the capability of mass-selectively minimizing matrix effects from salts, small molecules, and macromolecules, ultrasensitive detection of cytochrome C (>500-fold sensitivity improvement) in raw urine samples was demonstrated in less than 20 min.
W
and organic matrices) could be mass-selectively decreased. With minimum sample pretreatment, quantitative analyses of organic compounds, peptides, and proteins were achieved for samples with salt concentrations at physiological levels. Using dialysis membranes coated with antibodies, fast targeted protein quantitation (cytochrome C) in raw urine samples was achieved at the subppb level, >500-fold sensitivity improvement compared with conventional nanoESI. In comparison, high sensitivity detections (ppb level) have been reported for PS, but mostly for small molecules, such as drugs.37−40 PS has not been used to ionize proteins until 2014, in which year, Y. Zhang et al. investigated the structures of noncovalent protein complexes ionized by PS.38,46 The vertical separation capability of MESI was also demonstrated for the detection of multiple target molecules using a dual-layer membrane setup.
ith high sensitivity and specificity, mass spectrometry (MS) has been widely used in chemical identifications and detections. When coupled with separation techniques (liquid chromatography, LC, for example), it allows qualitative and quantitative analyses of trace amount analytes, even within complex backgrounds.1−5 Recent developments of ambient ionization techniques, including desorption electrospray ionization (DESI),6 direct analysis in real-time (DART),7 paper spray ionization (PS),8 and many others,9−22 permit direct sample analyses with minimized sample preparation and/ or separation processes. Nevertheless, sensitivity and quantitation accuracy of both LC−MS and ambient ionization-MS techniques can be substantially affected by the matrix effects,23,24 especially for biological samples with both organic (e.g., sugars, amines, and pigments) and inorganic (e.g., salts) matrixes. Efforts have been made for online removal of salts in electrospray ionization (ESI) using methods such as online membrane cleanup,25−28 adding solution additives,29 and etc.30,31 A number of ambient ionization methods have also been developed for analysis of biomolecules within highly concentrated salt solutions.32−36 Following the idea of PS, this study presents membrane electrospray ionization (MESI), which could achieve biomolecule analysis in raw biofluids with high sensitivity. As a unique nature of MESI, both vertical and horizontal chemical separations could be achieved in real-time. Matrix effects from interferences not only at the high-mass end (macromolecules and cells, for instance) but also at the low-mass end (inorganic © XXXX American Chemical Society
■
EXPERIMENTAL SECTION All experiments were carried out using a Bruker HCT mass spectrometer (Bruker Daltonics Inc., MA, Germany). Cytochrome C, myoglobin, lysozyme, angiotensin II, peptide MetArg-Phe-Ala (MRFA), progesterone, and urea were purchased from Sigma-Aldrich (MO). Cytochrome C antibody (catalog no. MAB898) was purchased from R&D systems (MN). Inorganic compounds and filter paper were purchased from Received: November 25, 2014 Accepted: March 1, 2015
A
DOI: 10.1021/acs.analchem.5b00467 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry
Figure 1. (a) On-membrane matrix removal and analyte ionization in MESI. (b) Scanning electron microscopy photos of the membrane before and after MESI.
Figure 2. Mass spectra of 10 μg mL−1 MRFA in PBS (100 mM, pH 7.4) (a) using MESI and (b) using nanoESI. Mass spectra of 500 μg mL−1 cytochrome C in 0.85% NaCl (c) using MESI and (d) using nanoESI. SNR comparisons between MESI and nanoESI for (e) 10 μg mL−1 MRFA and (f) 500 μg mL−1 cytochrome C with respect to different salt concentrations. Linear detection ranges for (g) MRFA in 0.85% NaCl and (h) cytochrome C in 0.85% NaCl. Membranes with MWCO 1 kDa were used for MRFA and 3.5 kDa for cytochrome C.
kV) and flushing solvent (methanol/water 6:4 v/v, 0.1% formic acid, 8 μL) are applied onto the membrane. MESI operation condition optimization process could be found in the Supporting Information. Molecules remaining on the membrane are transported and ionized at the tip of the triangular membrane. Because of the limited amount of sample and solvent loaded on the membrane, ion current could last ∼1 min (Figure S2 in the Supporting Information). The mass spectrum at the peak tops of the extracted ion chromatogram (EIC, m/z value of the target molecule was extracted) of MESI were chosen and used in this work. Figure 1b shows the scanning electron microscopy images on the membrane before and after flushing, in which almost all samples (progesterone in saline) were depleted after being flushed.
Beijing Chemical Company (Beijing, China). Dialysis membranes were purchased from Ebioeasy Co., Ltd. (Shanghai, China). Human blood, tear, urine, and saliva were provided by healthy volunteers in accordance with the requirements of medical ethics. In a MESI experiment (shown in Figure 1a and Figure S1 in the Supporting Information), a 2 μL sample is first loaded on a piece of pretreated triangular dialysis membrane (∼7 mm equilateral triangle, and details about the pretreatment of the membrane could be found in the Supporting Information), which was tightly covered on a piece of prewetted filter paper. The tip of the membrane is placed about 5 mm away from the MS inlet, and there is a 10−20° angle between the membrane surface and the horizontal plane to facilitate the movement of liquid on the membrane. In about 30 s, a high dc voltage (∼3.5 B
DOI: 10.1021/acs.analchem.5b00467 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry
Figure 3. Direct analyses of (a) cytochrome C in human tears, (b) angiotensin II in urine, (c) MRFA in human serum, and (d) progesterone in human saliva. Membranes with MWCO 1 kDa were used for MRFA and progesterone and 3.5 kDa for cytochrome C and angiotensin II (see text for details).
■
Because of the charge competition effects in ESI sources,41 the MS sensitivity and signal-to-noise ratio (SNR) would be substantially affected by the presence of nonvolatile organic and inorganic salts. With the help of a dialysis membrane, MESI could effectively diminish salts and achieve both high sensitivities and high SNRs for the MS analyses of biomolecules. As an example, Figure 2 compares the performances of MESI and nanoESI when analyzing MRFA and cytochrome C in solutions with salts. Working conditions of the nanoESI used in this study were optimized independently in terms of applied voltage and distance, which were then chosen to be ∼1 kV and 7 mm, respectively after optimization. Presented in Figure 2a−d, MESI could effectively improve the SNRs of mass spectra by reducing the noises from salts, thus minimizing the charge competition effects. The SNR improvements at different salt concentrations and for different analytes were plotted in parts e and f of Figure 2. Generally, the SNR enhancement is more prominent for samples with higher salt concentrations and larger analytes. About 10−20-fold SNR enhancements were observed for cytochrome C in NaCl solutions (1−5%). As shown in Figure 2g,h, the limit of quantitation (LOQ) for MRFA and cytochrome C in 0.85% NaCl solution were 10 ng mL−1 and 500 ng mL−1, respectively, when using MESI. These LOQ values were 10-fold lower than those obtained from nanoESI. Results for other compounds, as well as direct comparisons with PS, could be found in the Supporting Information (Figures S9 and S10). These results indicate that MESI offers a wider linear range and lower LOQ, especially for peptides or proteins whose molecular masses were relatively large (compared with the molecular weights of salts). A possible reason is membrane with larger pore sizes could be applied for larger targeted molecules, which could enhance the removal of salts and small interference molecules in the samples (Figure S4 in the Supporting Information).
RESULTS AND DISCUSSION The ionization mechanism of MESI was similar to that of PS, which was assisted by the high electric field generated at the tip of the triangle. However, the sample separation process in MESI is different from that of PS. In MESI, both vertical and horizontal chemical separations would happen at the same time, and ions were separated vertically based on their physical sizes. After samples being loaded on the membrane, salts and molecules with sizes smaller than the pore size of the membrane (or molecular weight cutoff, MWCO) could diffuse through the membrane and be absorbed by the filter paper. Figure S3 in Supporting Information monitored the dynamic salt removing process in a MESI, suggesting salts could be effectively removed in ∼15 s after applying the flushing solvent. Conventionally, it takes hours to perform a dialysis experiment. In a MESI experiment, the filter paper, gravity, electric driven force, and membranes with large pore sizes are believed to help the diffusion of small molecules through membranes in a much shorter period. By selecting the dialysis membrane with different MWCOs, different molecules could be massselectively retained on top of the membrane, as shown in Figure S4 in the Supporting Information. Optimization of the operation procedures and selection of different membranes for different targeted molecules could be found in the Supporting Information (Figures S5−S8). With cellulose-based microstructures, PS would have better separation effects horizontally, especially for matrix/interference molecules with high masses,8 which could not be filtered by membranes in the vertical direction. Furthermore, since the membrane is hydrophobic, membrane with a much larger MWCO was used to help the spread and transportation of solutions. The prewetted filter paper attached to the membrane also helps in terms of connecting liquid on the membrane with the high voltage, especially after the liquid on the membrane moved forward due to the consumption of spray and evaporation. C
DOI: 10.1021/acs.analchem.5b00467 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry
Figure 4. Ultrasensitive cytochrome C detection in human urine in 20 min: (a) MESI methodology using membranes coated with antibody; (b) the corresponding linear detection range of cytochrome C; (c) mass spectrum of 10 ng mL−1 (LOQ) cytochrome C in urine detected by MESI coated with anticytochrome C; (d) mass spectrum of 500 μg mL−1 cytochrome C in urine detected by MESI without anticytochrome C.
Figure 5. MESI with dual membrane layers for multiple target biomolecule analyses. (a) Diagram of MESI with dual membrane. (b) Cytochrome C detected from the first layer. (c) MRFA detected from the second layer and (d) the corresponding tandem mass spectrum.
With remarkable tolerance to matrix effects, MESI could be applied to the direct analyses of biofluids with minimum or no sample preparations. The detection of cytochrome C in human tears, angiotensin II in urine, MRFA in serum (diluted 10 times with water), and progesterone in human saliva could be
achieved with results shown in Figure 3. Cytochrome C and angiotensin II powders were directly spiked into human tears and urine, respectively. MRFA powder was spiked into 10-time diluted serum with water. A 200 μg mL−1 progesterone solution was first prepared in methanol and then mixed with human D
DOI: 10.1021/acs.analchem.5b00467 Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
■
saliva with a 1:1 ratio (v/v). A recovery rates between 97.4− 104.2% and a RSD < 2% were achieved for the biomolecules tested above (Table S1 in Supporting Information). In addition, MESI could also work in the negative mode as shown in Figure S11 in the Supporting Information. The detection sensitivities of targeted biomolecules could be further improved by coating the microdialysis membrane with antibodies. It was reported that cytochrome C in human urine can be utilized as a clinical biomarker for the diagnosis of druginduced liver injury.42 As shown in Figure 4a, the microdialysis membrane (MWCO 3.5 kDa) was first coated with anticytochrome C and then dipped into human urine samples spiked with cytochrome C at 37 °C for 15 min to allow the binding of cytochrome C with anticytochrome C. After washing 3 times by pure water, 10 μL 7 mol L−1 urea was added onto the membrane to separate antibody and antigen. Ionized using MESI, a LOQ of 10 ng mL−1 (500-fold improvement compared with 50 μg mL−1 using nanoESI in Figure 2h) and a linear quantitation range from 10 to 1250 ng mL−1 could be achieved (Figure 4b), which covers the entire clinical concentration range.42,43 Potentially, MESI coupled with antibody could also be applied to high-throughput antibody screening and specificity evaluations.44,45 Figure 4c,d showed mass spectra of cytochrome C detected by MESI coated with and without anticytochrome C in human urine, respectively, which suggested that immunoadsorption on the membrane can provide significant sensitivity improvement. Besides monitoring (or optimized for) one biomolecule, multiple biomolecules could be targeted at the same time using multiple layers of membranes with different pore sizes and pointing at different directions (Figure 5a). As an example, a mixture of cytochrome C (1 mg mL−1) and MRFA (100 μg mL−1) in 0.85% NaCl was loaded on a dual layer MESI source. A membrane with 3.5 kDa MWCO was chosen as the first layer, which stops cytochrome C; while the second layer has a MWCO of 1 kDa, which would retain MRFA. By facing the first membrane toward a MS, cytochrome C could be detected after flushing solvent onto the first layer (Figure 5b). By removing the first layer and turning the second membrane toward the MS, MRFA could then be flushed and detected (Figure 5c,d).
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Phone: +86-10-68918123. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We are grateful for the suggestions from Prof. Graham R. Cooks of Purdue University. This research was supported by NNSF China (Grant 21205005, 81471919), MOST China (Grant 2011YQ0900502), 1000 plan, Research Foundation of China CDC (Grant 2014A101).
■
REFERENCES
(1) Boggio, K. J.; Obasuyi, E.; Sugino, K.; Nelson, S. B.; Agar, N. Y.; Agar, J. N. Expert Rev. Proteomics 2011, 8, 591−604. (2) Date, S.; Mizuno, H.; Tsuyama, N.; Harada, T.; Masujima, T. Anal. Sci. 2012, 28, 201−203. (3) Mellors, J. S.; Jorabchi, K.; Smith, L. M.; Ramsey, J. M. Anal. Chem. 2010, 82, 967−973. (4) Dimri, G. P.; Lee, X.; Basile, G.; Acosta, M.; Scott, G.; Roskelley, C.; Medrano, E. E.; Linskens, M.; Rubelj, I.; Pereira-Smith, O. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 9363−9367. (5) Nemes, P.; Knolhoff, A. M.; Rubakhin, S. S.; Sweedler, J. V. Anal. Chem. 2011, 83, 6810−6817. (6) Takáts, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306, 471−473. (7) Cody, R. B.; Laramée, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 2297−2302. (8) Wang, H.; Liu, J.; Cooks, R. G.; Ouyang, Z. Angew. Chem., Int. Ed. 2010, 49, 877−880. (9) Venter, A.; Nefliu, M.; Cooks, R. G. TrAC, Trends Anal. Chem. 2008, 27, 284−290. (10) Westphall, M. S.; Jorabchi, K.; Smith, L. M. Anal. Chem. 2008, 80, 5847−5953. (11) Chetwani, N.; Cassou, C. A.; Go, D. B.; Chang, H. Anal. Chem. 2011, 83, 3017−3023. (12) Na, N.; Zhao, M.; Zhang, S.; Yang, C.; Zhang, X. J. Am. Soc. Mass. Spectrom. 2007, 18, 1859−1862. (13) Wei, Z.; Han, S.; Gong, X.; Zhao, Y.; Yang, C.; Zhang, S.; Zhang, X. Angew. Chem., Int. Ed. 2013, 52, 11025−11028. (14) Gong, X.; Zhao, Y.; Cai, S.; Fu, S.; Yang, C.; Zhang, S.; Zhang, X. Anal. Chem. 2014, 86, 3809−3816. (15) Vaikkinen, A.; Shrestha, B.; Kauppila, T. J.; Vertes, A.; Kostiainen, R. Anal. Chem. 2012, 84, 1630−1636. (16) Meier, L.; Berchtold, C.; Schmid, S.; Zenobi, R. Anal. Chem. 2012, 84, 2076−2080. (17) Haapala, M.; Suominen, T.; Kostiainen, R. Anal. Chem. 2013, 85, 5715−5719. (18) Liu, P.; Forni, A.; Chen, H. Anal. Chem. 2014, 86, 4024−4032. (19) Gilbert-Lopez, B.; Schilling, M.; Ahlmann, N.; Michels, A.; Hayen, H.; Molina-Díaz, A.; García-Reyes, J. F.; Franzke, J. Anal. Chem. 2013, 85, 3174−3182. (20) Zhan, X.; Zhao, Z.; Yuan, X.; Wang, Q.; Li, D.; Xie, H.; Li, X.; Zhou, M.; Duan, Y. Anal. Chem. 2013, 85, 4512−4519. (21) Gong, M.; Zhang, P.; MacDonald, B. D.; Sinton, D. Anal. Chem. 2014, 86, 8090−8097. (22) Wang, B.; Trimpin, S. Anal. Chem. 2014, 86, 1000−1006. (23) Taylor, P. J. Clin. Biochem. 2005, 38, 328−334. (24) King, R.; Bonfiglio, R.; Fernandez-Metzler, C.; Miller-Stein, C.; Olah, T. J. Am. Soc. Mass Spectrom. 2000, 11, 942−950. (25) Tibavinsky, I. A.; Kottke, P. A.; Fedorov, A. G. Anal. Chem. 2015, 87, 351−356. (26) Chen, Y.; Mori, M.; Pastusek, A. C.; Schug, K. A.; Dasgupta, P. K. Anal. Chem. 2011, 83, 1015−1021. (27) Sun, L.; Duan, J.; Tao, D.; Liang, Z.; Zhang, W.; Zhang, L.; Zhang, Y. Rapid Commun. Mass Spectrom. 2008, 22, 2391−2397.
■
CONCLUSIONS In summary, membrane electrospray ionization was introduced as a novel ambient ionization technique, and MESI-MS has satisfactory performances for targeted biomarker quantitation in raw biofluids. This technique shows great sensitivity for various biomolecule analyses, especially for large biomolecules (peptides or proteins) in very complex matrixes. With the capability of quantifying trace amount biomarkers in raw biofluids in real-time, MESI could accelerate the translation of mass spectrometry techniques to clinical applications.
■
Letter
ASSOCIATED CONTENT
S Supporting Information *
Materials and methods, investigation of the total ion chromatogram, selection of dialysis membranes, optimization of experimental conditions, comparison of MESI with nanoESI and PS, methodology validation, negative mode, and MESI using membrane coated with antibody. This material is available free of charge via the Internet at http://pubs.acs.org. E
DOI: 10.1021/acs.analchem.5b00467 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry (28) Chung, Y. T.; Ling, Y. C.; Yang, C. S.; Sun, Y. C.; Lee, P. L.; Lin, C. Y.; Hong, C. C.; Yang, M. H. Anal. Chem. 2007, 79, 8900−8910. (29) Mridul, K. M.; Lee, C. C.; Yutaka, H.; Zhan, Y.; Kenzo, H. Anal. Methods 2010, 2, 1905−1912. (30) Cheng, C.; Chen, C. S.; Hsieh, P. H. J. Chromatogr. A 2010, 1217, 2104−2110. (31) Ji, X.; Li, D.; Li, H. Biomed. Chromatogr. 2015, DOI: 10.1002/ bmc.3418. (32) Chang, D. Y.; Lee, C. C.; Shiea, J. Anal. Chem. 2002, 74, 2465− 2469. (33) Han, F.; Yang, Y.; Ouyang, J.; Na, N. Analyst 2015, 140, 710− 715. (34) Jackson, A. U.; Talaty, N.; Cooks, R. G.; Van Berkel, G. J. J. Am. Soc. Mass Spectrom. 2007, 18, 2218−2225. (35) Chen, H.; Venter, A.; Cooks, R. G. Chem. Commun. (Cambridge, U. K.) 2006, 19, 2042−2044. (36) Chen, H.; Yang, S.; Li, M.; Hu, B.; Li, J.; Wang, J. Angew. Chem., Int. Ed. 2010, 49, 3053−3056. (37) Lin, C.; Liao, W.; Chen, H.; Kuo, T. Bioanalysis 2014, 6, 199− 208. (38) Zhang, Y.; Ju, Y.; Huang, C.; Wysocki, V. H. Anal. Chem. 2014, 86, 1342−1346. (39) Wang, H.; Manicke, N. E.; Yang, Q.; Zheng, L.; Shi, R.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2011, 83, 1197−1201. (40) Wang, H.; Ren, Y.; McLuckey, M. N.; Manicke, N. E.; Park, J.; Zheng, L.; Shi, R.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2013, 85, 11540−11544. (41) Tang, K.; Page, J. S.; Smith, R. D. J. Am. Soc. Mass Spectrom. 2004, 15, 1416−1423. (42) Miller, T. J.; Knapton, A.; Adeyemo, O.; Noory, L.; Weaver, J.; Hanig, J. P. J. Appl. Toxicol. 2008, 28, 815−828. (43) Small, D. M.; Gobe, G. C. Expert Opin. Drug Metab. Toxicol. 2012, 8, 655−664. (44) Janda, K. D.; Schloeder, D.; Benkovic, S. J.; Lerner, R. A. Science 1988, 241, 1188−1191. (45) Roberts, S.; Cheetham, J. C.; Rees, A. R. Nature 1987, 328, 731−734. (46) Liu, J.; Wang, H.; Manicke, N. E.; Lin, J. M.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2010, 82, 2463−2471.
F
DOI: 10.1021/acs.analchem.5b00467 Anal. Chem. XXXX, XXX, XXX−XXX
