Ionization for Affinity Mass

A widely used protein array system is surface enhanced laser desorption/ ionization mass spectrometry (SELDI-MS),2 which is commercially available wit...
0 downloads 0 Views 132KB Size
Download PDF
Anal. Chem. 2004, 76, 3037-3041

Aptamer-Enhanced Laser Desorption/Ionization for Affinity Mass Spectrometry Lawrence W. Dick, Jr. and Linda B. McGown*

Department of Chemistry, Duke University, Box 90346, Durham, North Carolina 27708

The thrombin-binding DNA aptamer was used for affinity capture of thrombin in MALDI-TOF-MS. The aptamer was covalently attached to the surface of a glass slide that served as the MALDI surface. Results show that thrombin is retained at the aptamer-modified surface while nonspecific proteins, such as albumin, are removed by rinsing with buffer. Upon application of the low-pH MALDI matrix, the G-quartet structure of the aptamer unfolds, releasing the captured thrombin. Following TOF-MS analysis, residual matrix and protein can be washed from the surface, and buffer can be applied to refold the aptamers, allowing the surface to be reused. Selective capture of thrombin from mixtures of thrombin and albumin and of thrombin and prothrombin from human plasma was demonstrated. This simple approach to affinity capture, isolation, and detection holds potential for analysis, sensing, purification, and preconcentration of proteins in biological fluids. Recent advances in matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) have led to successful protein array technology based on a variety of interactions1. A widely used protein array system is surface enhanced laser desorption/ ionization mass spectrometry (SELDI-MS),2 which is commercially available with many options for surface substrates, including antibodies, metals, hydrophobic phases, lectins, enzymes, and receptors. These protein arrays have been highly successful in profiling proteomic mixtures and finding specific biomarkers for some common diseases.1,3,4 In the present work, we demonstrate the use of an aptamer substrate for affinity MALDI. Aptamers are oligonucleotides that exhibit high binding affinity for specific target molecules. They are combinatorially selected in vitro by means of an exponential enrichment process known as SELEX.5,6 Aptamers have been identified for dozens of target molecules, including proteins, enzymes, antibodies, peptides, and a host of small molecules, such as dyes, drugs, and amino acids. In chemical analysis, aptamers * Corresponding author. Phone: (919) 660-1545. Fax: (919) 660-1605. E-mail: [email protected]. (1) Tang, N.; Tornatore, P.; Weinberger, S. R. Mass Spectrom. Rev. 2004, 23, 34. (2) Hutchens, T. W.; Yip, T. T. Rapid Commun. Mass Spectrom. 1993, 7, 576. (3) Isaaq, H. J.; Veenstra, T. D.; Conrads, T. P.; Felschow, D. Biochem. Biophy.s Res. Commun. 2002, 292, 587. (4) Merchant, M.; Weinberger, S. Electrophoresis 2000, 21, 1164. (5) Tuerk, C.; Gold, L. Science 1990, 249, 505. (6) Ellington, A. D.; Szostak, J. W. Nature 1990, 346, 818. 10.1021/ac049860e CCC: $27.50 Published on Web 04/28/2004

© 2004 American Chemical Society

rival antibodies as affinity binding reagents.7,8 They have been investigated for use in numerous applications, such as surface staining,9 clinical diagnostics,10,11 sensing and detection,12-19 immunoassay,20 enzymatic analysis,21,22 combinatorial screening,23 and protein purification.24 The use of aptameric substrates for protein capture and analysis in affinity MALDI has been suggested,8 but not, to our knowledge, previously described in the published literature. Potential advantages that aptamers may offer over other affinity capture reagents include reusability, ease of production and manipulation, stability of DNA or modified oligonucleotides over a broad range of experimental conditions, and smaller size to facilitate higher surface coverage.25 This paper describes the use of an aptamer-modified surface for direct detection of thrombin via surface enhanced affinity MALDI-TOF-MS. The DNA thrombin-binding aptamer26-28 was covalently bound to a fused-silica glass surface. The aptamer adopts a G-quartet conformation in solution27,28 that plays a critical role in its interaction with thrombin. The presence of the G-quartet structure in the aptamer when it is covalently bound to a silica (7) Jayasena, S. D. Clin. Chem. 1999, 45, 1628. (8) Jenkins, R. E.; Pennington, S. R. Proteomics 2001, 1, 13. (9) Kawazoe, N.; Ito, Y.; Imanishi, Y. Anal. Chem. 1996, 68, 4309. (10) Brody, E. N.; Willis, M. C.; Smith, J. D.; Jayasena, S.; Zichi, D.; Gold, L. Mol. Diagn. 1999, 4, 381. (11) Golden, M. C.; Collins, B. D.; Willis, M. C.; Koch, T. H. J. Biotechnol. 2000, 81, 167. (12) Potyrailo, R. A.; Conrad, R. C.; Ellington, A. D.; Hieftje, G. M. Anal. Chem. 1998, 70, 3419. (13) German, I.; Buchanan, D. D.; Kennedy, R. T. Anal. Chem. 1998, 70, 4540. (14) Lee, M.; Walt, D. R. Anal. Biochem. 2000, 282, 142. (15) Srisawat, C.; Engelke, D. R. RNA 2001, 7, 632. (16) Stojanovic, M. N.; de Prada, P.; Landry, D. W. J. Am. Chem. Soc. 2001, 123, 4928. (17) Yamamoto, R.; Baba, T.; Kumar, P. K. Genes Cells 2000, 5, 389. (18) Frauendorf, C.; Jaschke, A. Bioorg. Med. Chem. 2001, 9, 2521. (19) Hamaguchi, N.; Ellington, A. D.; Stanton, M. Anal. Biochem. 2001, 294, 126. (20) Rye, P. D.; Nustad, K. Biotechniques 2001, 30, 290. (21) Drolet, D. W.; Moon-McDermott, L.; Romig, T. S. Nat. Biotechnol. 1996, 14, 1021. (22) Kato, T.; Yano, K.; Ikebukuro, K.; Karube, I. Analyst 2000, 125, 1371. (23) Green, L. S.; Bell, C.; Janjic, N. Biotechniques 2001, 30, 1094. (24) Romig, T. S.; Bell, C.; Drolet, D. W. J. Chromatogr., B 1999, 731, 275. (25) McGown, L. B.; Joseph, M. J.; Pitner, J. B.; Vonk, G. P.; Linn, C. P. Anal. Chem. 1995, 67, 663A. (26) Bock, L. C.; Griffin, L. C.; Latham, J. A.; Vermass, E. H.; Toole, J. J. Nature 1992, 355, 564. (27) Macaya, R. F.; Schultze, P.; Smith, F. W.; Roe, J. A.; Feigon, J. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 3745. (28) Wang, K. Y.; McCurdy, S.; Shea, R. G.; Swaminathan, S.; Bolton, P. H. Biochemistry 1993, 32, 1899.

Analytical Chemistry, Vol. 76, No. 11, June 1, 2004 3037

Figure 1. The thrombin binding aptamer in its G-Quartet conformation (See ref 26).

surface has been inferred in prior work,29-33 but thrombin binding to a thrombin-binding aptamer-modified fused-silica surface has not been previously demonstrated. Two different DNA sequences were used in this work, the thrombin-binding aptamer (5′-GGTTGGTGTGGTTGG-3′) and a scrambled oligonucleotide (5′- GGTGGTGGTTGTGGT-3′). The structure of the thrombin-binding aptamer27 is shown in Figure 1. The scrambled oligonucleotide does not form a G-quartet and does not bind to thrombin in solution;26 therefore, the scrambled sequence serves as a control to determine the specificity of the interactions of the thrombin-binding aptamer-modified surface with thrombin. MATERIALS AND METHODS Reagents. Human R-thrombin was obtained from Haematologic Technologies, Inc. (Essex Junction, VT) and was stored at -4 °C in 50% glycerol. Human serum albumin (HSA), 99% pure as determined by the manufacturer using agarose gel electrophoresis, was obtained from Sigma (St. Louis, MO) and stored at 4 °C. Stock solutions of 100 µM protein were prepared in phosphate buffer. The 5′-thiol modified oligonucleotides were custom-synthesized by Midland Certified Reagent Company (Midland, TX). The oligonucleotides were reconstituted in TrisHCl buffer, pH 7.2, and stored at -4 °C. Potassium phosphate and 3-aminopropyltriethoxysilane (3-APTES) were obtained from Sigma. Sulfosuccinimidyl-4-(N-maleimidomethyl)-cylcohexane-1carboxylate (SMCC) and tris(2-carboxyethyl)phosphine (TCEP) were obtained from Pierce Chemical (Rockford, IL). The MALDI matrix sinapinic acid (SA) was obtained from Sigma and made to 10 mg/mL in 30% acetonitrile in water with 0.1% trifluoroacetic acid. Pooled plasma (lyophilized) was obtained from Sigma and stored at 4 °C prior to reconstitution and -4 °C after reconstitution. Preparation of Oligonucleotide-Coated Fused-Silica Surfaces. Oligonucleotide-coated spots were prepared on a fusedsilica slide of 1 mM thickness (Valley Design Co., Westford, MA) using an adaptation of a previously described, covalent attachment method.34,35 In brief, the surface of the glass slide was activated (29) Kotia, R. B.; Li, L.; McGown, L. B. Anal. Chem. 2000, 72, 827. (30) Charles, J. A. M.; McGown, L. B. Electrophoresis 2002, 23, 1599. (31) Rehder, M. A.; McGown, L. B. Electrophoresis 2001, 22, 3759. (32) Silinski, M. A.; McGown, L. B. J. Chromatogr. A 2003, 1008, 233. (33) Joyce, M. V. Ph.D. Dissertation, Duke University, 2003. (34) Phillips, T. M.; Chmielinska, J. J. Biomed. Chromatogr. 1994, 8, 242. (35) O’Donnell, M. J.; Tang, K.; Ko ¨ster, H.; Smith, C. L.; Cantor, C. R. Anal. Chem. 1997, 69, 2438.

3038 Analytical Chemistry, Vol. 76, No. 11, June 1, 2004

Figure 2. Results for MALDI-TOF-MS of thrombin. Top: thrombin (3 pmol) on a conventional stainless steel MALDI plate. Bottom: aptamer spot incubated with 50 pmol of thrombin and rinsed.

by rinsing with methanol, water, and sodium hydroxide. Next, 3-APTES was attached to the activated silica surface. Spots of ∼2 mm diameter on the silated glass surface were then coated with the heterobifunctional linker. The 5′-thiol modified oligonucleotides were treated with TCEP to cleave the disulfide bonds and then reacted with the APTES linker on the glass surface to create the DNA-coated spots. The slides were then dried, sealed, and stored at room temperature. Protein Capture and Analysis. Proteins were incubated on the oligonucleotide-coated spots at room temperature. Following incubation, the glass chips were rinsed with 25 mM pH 8 potassium phosphate buffer (referred to hereinafter as “rinse buffer”) to remove any unbound or weakly bound species and dried. MALDI matrix (1.2 µL of SA) was added to the spots and allowed to crystallize. The glass chips were then mounted directly onto a conventional stainless steel MALDI target plate using double-sided tape. The spots were analyzed by TOF-MS using an Applied Biosystems Voyager DE. RESULTS AND DISCUSSION Figure 2 shows MALDI-TOF-MS results for thrombin recovered from a plain, stainless steel surface and from an aptamercoated spot. For the stainless steel surface, 3 pmol of thrombin was applied to the surface and analyzed, without incubation or rinsing, following application of the MALDI matrix. The aptamercoated spot was incubated with 50 pmol of thrombin for 30 min and then rinsed with buffer to remove unbound protein, followed by application of matrix and analysis. The two surfaces yielded

Figure 3. Results for an aptamer spot (top) and a scrambled spot (bottom), each incubated with 50 pmol of thrombin and rinsed.

comparable spectra, with somewhat higher signals for the stainless steel surface that may be due to a higher amount of protein at the surface or to more efficient desorption/ionization. As shown in the figure, four peaks are detected: two for the +1 and +2 thrombin and two other peaks near m/z of 13 000 and 25 000. The latter peaks are present in all of the mass spectra of the commercial thrombin and are attributed to degradation products or impurities in the commercial protein preparation. The thrombin peaks appear at slightly higher m/z at the glass surface than at the stainless steel surface. This is an experimental artifact due to the additional height contributed by the glass slide that sits atop the stainless steel target and perhaps to differences in conductivity between the stainless steel and fused-silica substrates. No signal was detected at either surface in the absence of the SA matrix (not shown). Results for detection of thrombin at the aptamer-coated and scrambled oligonucleotide-coated surfaces are shown in Figure 3. The experimental procedure was the same as described above for the aptamer-coated spot in Figure 2. The aptamer spots generally yielded signals that were at least 10 times higher than any signals observed for the scrambled oligonucleotide spots, indicating the affinity capture of thrombin by the thrombin-binding aptamer. The very low amounts of thrombin that were detected at some scrambled oligonucleotide spots may be due to nonspecific interactions of the protein with the DNA, protein adsorption to exposed fused-silica surfaces, or incomplete rinsing of unbound protein. The reusability of the aptamer spots was next examined. Results for the first use of an aptamer spot are shown in Figure 4A. After the first use, the spot was rinsed with 30% acetonitrile in water to remove the crystalline matrix as well as thrombin that would be released due to unfolding of the G-quartet structure upon application of the low-pH matrix solution (unfolding of the

Figure 4. Results for reuse of an aptamer spo: (A) results for first use; (B) results after first use, following 30% acetonitrile rinse and regeneration with buffer; (C) results for second use. For first use (A) and second use (C), 50 pmol of thrombin was incubated on the spot for 30 min at room temperature, followed by a buffer rinse and application of SA matrix. No additional protein was added between (A) and (B).

structure at pH 2 was verified from circular dichroism spectra of bulk solutions of the aptamer). The plates then were rinsed with phosphate buffer and allowed to dry, followed by application of the matrix solution and MALDI-TOF-MS analysis. As shown in Figure 4B, no significant protein was detected, indicating that the acetonitrile rinse removed any thrombin that remained on the aptamer spot after the first use. Figure 4C shows the results for the second use of the same aptamer spot, which are similar to those obtained in the first use. These results are representative of results obtained for several aptamer spots. They demonstrate the reusability of the aptamer spots for affinity protein capture and MALDI-TOF-MS detection. Figure 5 shows the MALDI-TOF-MS results for spots that were incubated with different amounts of thrombin, ranging from 5 to 50 pmol. The results show optimal detection at 10 pmol for the aptamer-coated surface, above which the protein concentration becomes too high for efficient desorption/ionization from the MALDI matrix. The detectability varied among aptamer spots, and signal was observed for as little as 5 pmol in the best case. Figure 6 shows signals for different amounts of thrombin at unmodified stainless steel and fused-silica surfaces. For both surfaces, optimal detection occurred around 3-6 pmol. It should be noted that the system has not yet been optimized with respect to MALDI matrix composition or experimental conditions. Analytical Chemistry, Vol. 76, No. 11, June 1, 2004

3039

Figure 5. Results for aptamer spot incubated with 50, 20, 10, and 5 pmol of thrombin and rinsed. Note change of scale for 5-pmol sample.

Figure 7. Results for an aptamer spot incubated with an equimolar mixture of thrombin and HSA (25 pmol each) for 4 h at room temperature and rinsed with buffer.

Figure 6. Results for 6, 3, and 1 pmol of thrombin at unmodified stainless steel (left) and fused-silica (right) surfaces. Surfaces were not rinsed after application of protein.

To examine the specificity of protein capture, the aptamer spots were incubated with an equimolar mixture of thrombin and HSA. As shown in Figure 7, there is a very small peak around m/z 66 000, corresponding to the molecular weight of HSA, in addition to the large peak from thrombin. Repeated trials using the same amount of HSA alone gave results ranging from no signal to peaks of the same small magnitude, as shown for the protein mixture in Figure 7. The HSA peak is most likely due to residual protein that is weakly associated with the aptamer or adsorbed to exposed fused-silica surfaces. The results demonstrate the high specificity of protein capture by the aptamer. Experiments were then conducted using pooled human plasma. Figure 8 (top) shows results for an aptamer spot incubated with 3040 Analytical Chemistry, Vol. 76, No. 11, June 1, 2004

the plasma. The appearance of a peak at the molecular weight corresponding to thrombin demonstrates selective capture of thrombin from the complex plasma matrix by the aptamermodified surface. The mass spectrum shows an additional peak near m/z 58 000, which corresponds to the mass of prothrombin. There are also peaks corresponding to the m/z values of HSA+1 and HSA+2. This is not surprising, since albumin is present at millimolar concentrations (i.e., a 1000-fold excess over prothrombin) in plasma. The recovery of thrombin from the plasma is particularly encouraging because the concentration of thrombin in plasma is expected to be low. The predominant, stable form of the protein is prothrombin, a zymogen that is cleaved to yield the activated R-thrombin upon appropriate signaling.36 The presence of any (36) Blomback, B.; Hanson, L. A. Plasma Proteins; John Wiley and Sons: New York, 1979.

fragments to thrombin proteins that might suppress ionization of the proteins. Figure 8 (bottom) shows the results for the pooled human plasma on a scrambled spot. The peaks that correspond to thrombin and prothrombin, which were observed at the aptamer spot, are not seen at the scrambled spot. Peaks from the singly and doubly charged albumin are present in both spectra, which again indicates nonspecific adsorption of the highly abundant albumin. Further optimization of the capture and rinse protocols is underway to reduce interference from nonspecific protein adsorption. CONCLUSIONS

Figure 8. Results for pooled human plasma. Top: aptamer spot incubated with 5 µL of plasma and rinsed. Middle: aptamer spot incubated with 5 µL of plasma spiked with 25 pmol of thrombin and rinsed. Bottom: scrambled spot incubated with 5 µL of plasma and rinsed.

detectable thrombin in its active (i.e., cleaved) form is likely due to stabilizers in the pooled plasma preparation. The plasma concentration of prothrombin is ∼90 mg/L,36 which corresponds to ∼1.5 µM, or 7.5 pmol on the aptamer spot. Figure 8 (middle) shows the results for pooled human plasma that was doped with the commercial thrombin. The spectrum exhibits two sharp peaks at low molecular weight (m/z 30004000), but no thrombin or prothrombin was detected. The low molecular weight peaks, which were absent from the spectrum in Figure 8 (top), may be due to formation of insoluble fibrin peptide fragments upon addition of the commercial thrombin to the plasma.36 In its active form, thrombin causes fibrinogen to form insoluble fibrin that might not be removed by the buffer rinse. The absence of thrombin or prothrombin peaks may be due to preferential desorption/ionization of the low molecular weight species in the MALDI matrix or to a high ratio of fibrin

The results demonstrate the affinity capture and detection of thrombin on a thrombin-binding aptamer-coated surface. Comparison with a surface coated with a scrambled oligonucleotide that cannot form a G-quartet clearly shows that the thrombin is captured by specific, strong affinity interactions. The capture mechanism can be attributed to the presence of the G-quartet conformation on the aptamer spot, providing confirmation of G-quartet formation by the covalently bound aptamer at the glass surface. This is an important result not only for the present work but also for previous studies of aptamer-coated capillaries in opentubular capillary electrochromatography, from which G-quartet formation was inferred but unproven.29-32 The studies show that nonspecific proteins, such as albumin, are largely removed from the aptamer spot simply by rinsing with buffer. The studies of plasma demonstrate selective capture of thrombin and prothrombin from the complex plasma matrix. The capture/release/analysis cycle was able to be repeated on the same spot, showing that the unfolding of the aptamer upon application of low pH matrix is reversible and that the spots can be reused. According to the results, the cycle proceeds as follows: thrombin is captured by the G-quartet aptamer, nonspecifically bound proteins are removed by rinsing the surface with buffer, the application of the matrix unfolds the aptamer to release the thrombin that is then detected by TOF-MS; a 30% acetonitrile rinse removes the matrix and any remaining thrombin from the surface, which is then reconstituted by rinsing with buffer, readying the surface for another experiment. This simple, aptamer-based approach to affinity-enhanced MALDI-MS has potential for use in proteomic analysis and affinitybased methods for sensing, purification, and preconcentration of proteins from biological fluids. Optimization studies and investigations of other aptamer systems are underway. ACKNOWLEDGMENT This work was supported by the National Institutes of Health (Grant 1R03 AG21742-01).

Received for review January 23, 2004. Accepted March 25, 2004. AC049860E Analytical Chemistry, Vol. 76, No. 11, June 1, 2004

3041