Anal. Chem. 2010, 82, 3686–3691
Matrix-Assisted Laser Desorption Ionization-Mass Spectrometry Signal Enhancement of Peptides after Selective Extraction with Polymeric Reverse Micelles Nadnudda Rodthongkum, Yangbin Chen, S. Thayumanavan, and Richard W. Vachet* Department of Chemistry, University of Massachusetts Amherst, Amherst, Massachusetts 01003 Extraction of peptides by reverse micelle-forming amphiphilic homopolymers and subsequent matrix-assisted laser desorption ionization-mass spectrometry (MALDIMS) detection of these peptides in the presence of these polymers can significantly enhance peptide ion signals. Here, the mechanism of this MALDI signal enhancement is investigated. We find that the signal enhancement is caused by coalescence of polymer-peptide conjugates into “hotspots” on the MALDI target. Hotspot formation is observed only on hydrophilic surfaces and not hydrophobic surfaces. With the use of an Anchorchip MALDI target, which contains very small hydrophilic spots surrounded by a larger hydrophobic area, we find that this hotspot formation can be further exploited for ultrasensitive MALDI-MS analyses of peptides and peptide mixtures. MALDI-MS signals can be enhanced by 3-5 orders of magnitude when peptides are extracted by the amphiphilic homopolymers and detected on the Anchorchip MALDI target. This signal enhancement combined with the extraction selectivity of these reverse micelle-forming homopolymers makes these materials promising tools for sensitive detection of peptides in complex mixtures. Analyses of peptides and proteins in complex biological samples (i.e., multiprotein digests, cell lysates) are an important part of biological, medicinal, and environmental studies.1-6 Because of its sensitivity, speed, and ability to provide structural information, mass spectrometry (MS) is a key technique used for these analyses.7-10 Even so, efficient and selective sample preparation techniques that can eliminate interferences while extracting and concentrating analytes of interest from complex * Corresponding author. Department of Chemistry, LGRT 701, 710 N. Pleasant St., University of Massachusetts Amherst, Amherst, MA 01003. E-mail: rwvachet@ chem.umass.edu. (1) Hanash, S. Nature 2003, 422, 226–232. (2) Decramer, S.; Wittke, S.; Mischak, H.; Zurbig, P.; Walden, M.; Bouissou, F.; Bascands, J. L.; Schanstra, J. P. Nat. Med. 2006, 12, 398–400. (3) Yanofsky, C. M.; Bell, A. W.; Lesimple, S.; Morales, F.; Lam, T. T.; Blankney, G. T.; Marshall, A. G.; Carrillo, B.; Lekpor, K.; Boismenu, D.; Kerney, R. E. Anal. Chem. 2005, 77, 7246–7254. (4) Kamysz, W.; Okroj, M.; Lempicka, E.; Ossowski, T.; Lukasiak, J. Acta Chromatogr. 2004, 14, 180–186. (5) Pan, C. S.; Xu, S. Y.; Zhou, H. J.; Fu, Y.; Ye, M. L.; Zou, H. F. Anal. Bioanal. Chem. 2007, 193–204. (6) Chaves, A. R.; Silva, S. M.; Queiroz, R. H. C.; Lancas, F. M.; Queiroz, M. E. C. J. Chromatogr., B 2007, 850, 295–302.
3686
Analytical Chemistry, Vol. 82, No. 9, May 1, 2010
mixtures are greatly desired to simplify the subsequent mass spectral analyses. The importance of such sample preparation methods is evidenced by the numerous approaches to this problem.11-20 Recently, we have investigated the extraction capabilities of a new type of amphiphilic homopolymer that can be effectively used in a two-phase liquid-liquid format.21 Two-phase liquid-liquid extractions have some inherent advantages because analyte extraction, purification, and concentration can be performed in a single step; however, these extractions are not generally useful for polar analytes, such as peptides and proteins, unless the organic phase is modified. To address this limitation, we use reverse-micelle forming amphiphilic homopolymers (I)22-24 to modify the organic phase so that positively charged peptides can be selectively extracted.21,25,26 The extracted peptides are then readily detected by matrix-assisted laser desorption ionizationmass spectrometry (MALDI-MS) with minimal sample workup. We have also found that by varying the aqueous phase pH during (7) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299–2332. (8) Belov, M. E.; Gorshkov, M. V.; Udseth, H. R.; Anderson, G. A.; Smith, R. D. Anal. Chem. 2000, 72, 2271–2279. (9) Keller, B. O.; Li, L. J. Am. Soc. Mass Spectrom. 2001, 12, 1055–1063. (10) Aebersold, R.; Mann, M. Nature 2003, 422, 198–207. (11) Zhang, H.; Yeung, K. K. C. Anal. Chem. 2004, 76, 6814–6818. (12) Wong, V. N.; Fernando, G.; Wagner, A. R.; Zhang, J.; Kinsel, G. R.; Zauscher, S.; Dyer, D. J. Langmuir 2009, 25, 1459–1465. (13) Silvertand, L. H. H.; Torano, J. F.; De Jong, G. J.; Van Bennekom, W. P. Electrophoresis 2009, 30, 1828–1835. (14) Nesbitt, C. A.; Yeung, K. K. C. Analyst 2009, 134, 65–71. (15) Wall, D. B.; Parus, S. J.; Lubman, D. M. J. Chromatogr., B 2002, 774, 53– 58. (16) Sudhir, P. R.; Shrivas, K.; Zhou, Z. C.; Wu, H. F. Rapid Commun. Mass Spectrom. 2008, 22, 3076–3086. (17) Wang, J.; Chen, R.; Ma, M.; Li, L. Anal. Chem. 2008, 80, 491–500. (18) Tempels, F. W. A.; Underberg, W. J. M.; Somsen, G. W.; Jong, G. J. D. Electrophoresis 2007, 28, 1319–1326. (19) Garaguso, I.; Borlak, J. Proteomics 2008, 8, 2583–2595. (20) Tyan, Y. C.; Yang, M. H.; Liao, P. C.; Liao, J. D.; Jong, S. B.; Liu, C. Y.; Wang, M. C.; Grunze, M. Int. J. Mass Spectrom. 2007, 262, 67–72. (21) Combariza, M. Y.; Savariar, E. N.; Vutukuri, D. R.; Thayumanavan, S.; Vachet, R. W. Anal. Chem. 2007, 79, 7124–7130. (22) Basu, S.; Vutukuri, D. R.; Shyamroy, S.; Sandanaraj, B. S.; Thayumanavan, S. J. Am. Chem. Soc. 2004, 126, 9890–9891. (23) Basu, S.; Vutukuri, D. R.; Thayumanavan, S. J. Am. Chem. Soc. 2005, 127, 16794–16795. (24) Savariar, E. N.; Aathimanikandan, S. V.; Thayumanavan, S. J. Am. Chem. Soc. 2006, 128, 16224–16230. (25) Gomez-Escudero, A.; Azagarsamy, M. A.; Theddu, N.; Vachet, R. W.; Thayumanavan, S. J. Am. Chem. Soc. 2008, 130, 11156–11163. (26) Rodthongkum, N.; Washington, J. D.; Savariar, E. N.; Thayumanavan, S.; Vachet, R. W. Anal. Chem. 2009, 81, 5046–5053. 10.1021/ac1000256 2010 American Chemical Society Published on Web 04/08/2010
the extraction, peptide mixtures can be fractionated in such a way that titration curves can be simultaneously obtained for all peptides in a mixture.26 We think that these titration curves, along with accurate mass measurements, represent a new way of improving protein identification methods.
Interestingly, we have also observed that selective peptide extraction with these amphiphilic polymers is accompanied by a significant MALDI signal enhancement when the peptides are analyzed in the presence of the polymer. In previous work, we demonstrated that this enhancement, along with the concentration factors associated with the extraction, allowed us to obtain reproducible ion signals for peptides at concentrations as low as 10 pM.21 In this article, we study the mechanism by which peptide ion signals are enhanced in the presence of the homopolymer with the goal of further improving this enhancement. Using fluorescence microscopy, we find that the amphiphilic homopolymer helps to organize peptides into “hotspots” on the MALDI target surface. Moreover, we find that hotspot formation is observed only on hydrophilic surfaces but not hydrophobic ones. We then exploit this surface-dependent hotspot formation to enhance MALDI-MS signals by 3-5 orders of magnitude. EXPERIMENTAL SECTION Reagents. Bradykinin peptide (MW 1060, RPPGFSPFR) and FITC-labeled bradykinin (MW 1563, FITC-C6-RPPGFSPFR), kinetensin (MW 1172, IARRHPYFL), ACTH human (MW 1623, SYSMEHFRWGKPV), angiotensin I (MW 1296, DRLVYIHPFHL), and spinorphin (MW 877, LVVYPWT) were purchased from American Peptide Company Sunnyvale, CA). Cytochrome c (horse heart), R-cyano-hydroxycinnamic acid (R-CHCA), tris(hydroxymethyl)aminomethane/tris(hydroxymethyl)aminomethane hydrochloride (Tris/Tris-HCl), chlorotrimethylsilane, toluene, methanol (MeOH), and trifluoroacetic acid (TFA) were obtained from Sigma-Aldrich (St. Louis, MO). Trypsin was acquired from Promega (San Luis Obispo, CA). Tetrahydrofuran (THF) was purchased from Fisher (Pittsburgh, PA) and then distilled over Na/Ph2CO before use. All other chemicals were used as provided. The water used in preparation of all solutions was obtained from a Milli-Q water purification system (Millipore, Bedford, MA). The polymer was prepared following the procedure previously reported, and the synthetic details are described elsewhere.22 Reverse Micelle Formation. The reverse micelle solution was prepared at a 1.0 × 10-4 M concentration by dissolving 10 mg of polymer in toluene and adding 2 mol equiv of water per monomer to form the water pool in the reverse micelle interiors. This solution was sonicated until a visibly clear solution was obtained (∼3 h) and then used for the liquid-liquid extraction. Proteolytic Digestion. Cytochrome c was digested by initially mixing 110 µL of the protein solution (∼3.0 × 10-4 M in 50 mM of Tris/1 mM of CaCl2) with 40 µL of MeOH and heated at 60
°C for 15 min in order to denature the protein. A total of 150 µL of a 4.4 µg/mL solution of trypsin, which contained 50 mM of Tris and 1 mM of CaCl2, was then mixed with denatured cytochrome c solution. The resulting solution was then incubated at 37 °C for 18 h. The digestion was stopped by filtering the solution through a 10 000 molecular weight cutoff (MWCO) Centricon filter and centrifuged at 12 000 rpm for 15 min. The filtrate was then kept in the freezer until the extraction and analysis were performed. Extraction Procedure. The two-phase liquid-liquid extraction protocol recently developed by our group was employed.21 An aqueous solution of peptide was prepared in 50 mM Tris/ Tris-HCl solution at the desired pH using 0.6 M HCl or NaOH for pH adjustment. The extraction involves mixing 200 µL of toluene solution of 1.0 × 10-4 M reverse micelles with 1 mL of an aqueous peptide solution (1.0 × 10-7 M), buffered to the desired pH. The mixture was vortexed and then centrifuged at 12 000 rpm for 30 min to separate the resulting emulsion into two phases. The organic phase was dried and redissolved in 10 µL of distilled THF. This solution was mixed with 20 µL of R-CHCA matrix solution (0.16 M in 60:40:0.3% THF/H2O/ TFA), for a final volume of 30 µL, and directly spotted on a MALDI target or glass slide using the dried droplet method. The dried sample from the organic phase was then analyzed either by MALDI-MS or fluorescence microscopy. The remaining aqueous phase was analyzed by MALDI-MS. For this experiment, equal amounts (5 µL) of the aqueous phase and the matrix solution (0.16 M in 60:40:0.3% THF/H2O/TFA) were mixed together, and 1 µL of this solution was spotted on a MALDI target. MALDI-MS Analysis. A Bruker Reflex III MALDI time-offlight mass spectrometer or a Bruker Omniflex MALDI time-offlight mass spectrometer was used to perform the MALDI-MS analyses. When using the Reflex III, all mass spectra were obtained in the reflectron mode using a 20 kV accelerating voltage with an average of 60 shots at 45% laser power. When the Omniflex is used, all mass spectra were acquired in the reflectron mode and represent an average of 100 shots acquired at 15% laser power; the accelerating voltage was set to 19 kV. One of three types of targets was used for the MALDI-MS analyses. For most experiments, a normal stainless steel MALDI target was used. For the MALDI analyses on the glass target, a glass microscope slide was attached to the instrument’s normal stainless target and inserted into the mass spectrometer. Bruker Daltonics’ Anchorchip MALDI target was used in the experiments that required small hydrophilic spots. For the Anchorchip MALDI target, 0.5 µL of the analyte/ matrix solution was carefully deposited onto the small hydrophilic center at the core of the sample spot in order to force the analyte to form “hotspots” on these areas. In most cases, the “hotspots” have dimensions greater than 50-75 µm, so they can be readily visualized by the microscope camera that is focused on the MALDI target. All MALDI mass spectra are acquired by focusing the laser on these readily observable “hotspots.” Modification of the Glass MALDI Target. A microscope glass slide was left in concentrated sulfuric acid for 3 h to oxidize the glass surface and was subsequently washed by Milli-Q water. This glass slide was then immersed in chlorotrimethylsilane overnight (∼12 h) at 37 °C to generate the hydrophobic coated Analytical Chemistry, Vol. 82, No. 9, May 1, 2010
3687
Figure 1. MALDI spectra of a 3.3 × 10-6 M nonextracted sample of bradykinin (left) and a 1.0 × 10-7 M reverse-micelle extracted sample of bradykinin (right). The S/N ratio and the error associated with the bradykinin ion signals are 700 ( 200 (left) and 5500 ( 500 (right), respectively. The reported S/N ratios and errors are averages and standard deviations obtained from 15 different spectra. These 15 spectra come from 3 different spots (and thus different samples) on the MALDI target and 5 replicates on each spot with the laser focused at different positions in each of the 5 replicates.
glass surface. The hydrophobic surface property was tested by measuring the contact angle of a water droplet added to the modified surface. The resulting value was then compared to the value of a droplet added to the unmodified glass surface. Contact Angle (CA) Measurements. The CA measurements were made with a Rame´-Hart telescopic goniometer and a Gilmont syringe with a 24-gauge flat-tipped needle. The probe fluid used was Milli-Q water. Dynamic advancing (θa) and receding angles (θr) were recorded while the probe fluid was added to and withdrawn from the drop, respectively. The values reported are averages of 3-5 measurements made on different areas of the modified glass surface. We used dynamic CA measurements to characterize the wettability of the sample glass slides. The modified surface becomes hydrophobic with θa/θr ) 78 ± 2°/59 ± 2°. In comparison, the unmodified glass slide is quite hydrophilic with θa/θr ) 22 ± 1°/9 ± 1°. Fluorescence Microscopy. Fluorescence microscopy, which is commonly used to study the distribution of peptides in MALDI matrix,27,28 was used to investigate the MALDI signal enhancement phenomenon. The microscopy experiments were performed on an Olympus IX51 fluorescence microscope using 20× resolving power. All experiments were performed in both white light and fluorescence modes (λem 515 nm). To prepare the sample for analysis, 1 µL of the sample/matrix solution was spotted on a 25 mm × 75 mm × 1 mm glass slide using the dried droplet method. RESULTS AND DISCUSSION MALDI Signal Enhancement of Bradykinin and FITCBradykinin. The carboxylate interiors of the polymeric reverse micelles cause positively charged peptides to be selectively extracted via Coulombic attraction, yet this extraction is accompanied by significant MALDI signal enhancement when the peptides are analyzed in the presence of the polymer. If we assume that 100% of a given peptide is extracted from the aqueous solution into the negatively charged reverse micelles, then a 33-fold increase in peptide concentration is expected. This 33-fold increase is due to the volume difference between aqueous and organic (27) Dai, Y.; Whittal, R. M.; Li, L. Anal. Chem. 1996, 68, 2494–2500. (28) Tummala, R.; Green-Church, K. B.; Limbach, P. A. J. Am. Soc. Mass Spectrom. 2005, 16, 1438–1446.
3688
Analytical Chemistry, Vol. 82, No. 9, May 1, 2010
phases in the two-phase liquid-liquid extraction (1 mL of the initial aqueous solution vs 30 µL of the final organic solution). Even considering for this 33-fold increase in peptide concentration, the MALDI signals for peptide ions are approximately 1 order of magnitude higher than expected. As an example, when a 1.0 × 10-7 M solution of the peptide bradykinin is extracted by polymeric reverse micelles, the resulting MALDI ion signal is about 8 times greater than expected (Figure 1). We arrive at this conclusion by comparing a MALDI spectrum of a 3.3 × 10-6 M solution of bradykinin (left), which is the expected peptide concentration after extraction, to a 1.0 × 10-7 M bradykinin solution that has been extracted by the polymeric reverse micelles (right). The (M + H)+ of bradykinin from the extracted sample has a signal-to-noise (S/N) ratio of about 5500, while the same ion from the higher concentration bradykinin sample, which was not extracted, has a S/N ratio of about 700. A possible reason for the observed signal differences in Figure 1 is that the presence of polymer during the MALDI-MS process enhances the peptide ion signal. We hypothesize that the presence of polymer in the matrix forces the peptides to organize on the MALDI target surface in a way that generates analyte-rich zones or “hotspots.” The peptide concentration would be increased in these hotspots, thereby enhancing the MALDI signal of the extracted peptide. To explore this hypothesis, we obtained fluorescence microscopy images of FITC-labeled bradykinin samples. Before taking these images, however, we first confirmed that the signal enhancement was still achieved with this labeled peptide on the glass surfaces used in the microscopy experiments. After extraction and MALDI-MS analysis of bradykinin on a glass sample target, we find that the signal enhancement (∼8-fold) is very similar to that on a stainless steel target (Figure S1 in the Supporting Information). Similarly, the MALDI ion signal for the extracted sample of FITC-labeled bradykinin on a glass target was also enhanced, although to a slightly lesser degree than unlabeled bradykinin (Figure S2 in the Supporting Information). For a range of peptides with 7-13 amino acids, we find that the MALDI signal enhancement due to the polymer extraction usually varies from a factor of 3 to as high as a factor of 10 (data not shown). Investigation of Signal Enhancement Mechanism by Fluorescence Microscopy. To test our hypothesis, nonextracted FITC-labeled bradykinin (3.3 × 10-6 M) and extracted FITC-
Figure 2. (a) Normal (left) and fluorescence (right) microscope images of nonextracted FITC-bradykinin (3.3 × 10-6 M) mixed with R-CHCA. (b) Normal (left) and fluorescence (right) microscope images of extracted FITC-bradykinin (1.0 × 10-7 M before extraction) mixed with R-CHCA. In the extracted samples, the polymer is still present.
labeled bradykinin (1.0 × 10-7 M before extraction) were separately mixed with a matrix solution of R-CHCA and applied to separate glass slides using the dried-droplet method. When dry, the matrix-peptide mixtures were immediately imaged with a fluorescence microscope. Figure 2 shows the normal and fluorescent images of the two samples. It is clear from Figure 2 that the extracted FITC-labeled peptides aggregate into hotspots, whereas the nonextracted FITClabeled peptides do not. This observation indicates that the polymer causes the peptides to coalesce together rather than cocrystallize more homogeneously throughout the matrix. This phenomenon probably occurs because the positively charged peptides are still associated with the negatively charged polymers rather than the matrix. We tentatively conclude that the resulting hotspots are the likely cause of the MALDI signal enhancement observed with the extracted peptides. To further examine if the peptide concentration associated with the hotspot formation is the main cause of the signal enhancement, the amount of light in different areas of the fluorescence images were determined for both the nonextracted and extracted samples. To avoid saturation of the microscope’s detector at any given spot, relatively low (
