Online Coupling of Electrochemical Reactions with Liquid Sample

Oct 29, 2009 - Lars Büter , Helene Faber , Tina Wigger , Martin Vogel , and Uwe Karst. Analytical Chemistry 2015 87 (19), 9931-9938. Abstract | Full ...
0 downloads 0 Views 2MB Size
Download PDF
Anal. Chem. 2009, 81, 9716–9722

Online Coupling of Electrochemical Reactions with Liquid Sample Desorption Electrospray Ionization-Mass Spectrometry Jiwen Li, Howard D. Dewald, and Hao Chen* Center for Intelligent Chemical Instrumentation, Department of Chemistry and Biochemistry, Clippinger Laboratories, Ohio University, Athens, Ohio 45701 The combination of electrochemistry (EC) and mass spectrometry (MS) is a powerful analytical tool to study redox reactions. This work reports the online coupling of a thin-layer electrochemical flow cell with liquid sample desorption electrospray ionization mass spectrometry (DESI-MS) and its applications in investigating various electrochemical reactions of biological molecules such as oxidative formation and reductive cleavage of disulfide bonds and online derivatization of peptides/proteins. As a result of the direct sampling nature of DESI, several useful features of such a coupling have been found, including simple instrumentation, fast response time (e.g., 3.6 s in the case of dopamine oxidation), freedom to choose a favorable ionization mode of DESI or traditional electrolysis solvent systems, and the absence of background signal possibly resulting from ionization when the cell is off (e.g., in the case of dopamine oxidation). More importantly, with the use of this new coupling apparatus, three disulfide bonds of insulin were fully cleaved by electrolytic reduction and both the A and B chains of the protein were successfully detected online by DESI-MS. In addition, online tagging of free cysteine residues of peptides/proteins employing electrogenerated dopamine o-quinone can be performed. These revealed characteristics of the coupling along with examined electrochemical reactions suggest that EC/DESI-MS has good potential in bioanalysis. The online combination of electrochemistry (EC) with mass spectrometry (MS) has proven useful in the identification of the products or intermediates of electrochemical reactions, leading to many applications in bioanalysis, as well as in mechanistic studies of redox reactions. The advantage of adopting MS as an EC detector stems from the fact that MS is sensitive and can provide molecular weight information and that tandem MS can be also used for structural analysis based on ion dissociation. The first EC/MS coupling was carried out by Bruckenstein1 using a nonwetting, porous membrane to allow the sampling of gaseous and volatile EC products by MS. Following this seminal work, * To whom correspondence should be addressed. Hao Chen, Assistant Professor, Department of Chemistry and Biochemistry, Clippinger Laboratories, Ohio University, Athens, OH, 45701. Phone: 740-331-2302. Fax: 740-597-3157. E-mail: [email protected]. (1) Bruckenstein, S.; Gadde, R. R. J. Am. Chem. Soc. 1971, 93, 793–794.

9716

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

coupling EC/MS to study nonvolatile solution species were further realized using such ionization methods as thermospray (TS),2-4 fast atom bombardment (FAB),5 and electrospray ionization (ESI).6-13 In particular, Van Berkel and other investigators have conducted significant pioneering research with regard to electrochemistry combined online with ESI-MS,6-13 as well as the inherent electrochemistry occurring in the ESI.14-16 The strength of EC/ESI-MS coupling is that ESI is powerful in ionizing nonvolatile, polar, thermally labile compounds and high mass biological samples and has tolerances to a wide range of solvent systems.17 Also, EC/ESI-MS has been used to mimic biologically relevant electrochemical reactions,18,19 for the conjugation with chromatographic separation,11 for online chemical tagging,20 and for oxidative cleavage of peptides/proteins.21,22 The analytical aspects of EC/ESI-MS have been reviewed.23,24 (2) Hambitzer, G.; Heitbaum, J. Anal. Chem. 1986, 58, 1067–1070. (3) Dewald, H. D.; Worst, S. A.; Butcher, J. A.; Saulinskas, E. F. Electroanalysis 1991, 3, 777–782. (4) Volk, K. J.; Yost, R. A.; Brajter-Toth, A. Anal. Chem. 1989, 61, 1709–1717. (5) Bartmess, J. E.; Phillips, L. R. Anal. Chem. 1987, 59, 2012–2014. (6) Zhou, F.; Van Berkel, G. J. Anal. Chem. 1995, 67, 3643–3649. (7) Bond, A. M.; Colton, R.; D’Agostino, A.; Downard, A. J.; Traeger, J. C. Anal. Chem. 1995, 67, 1691–1695. (8) Xu, X.; Lu, W.; Cole, R. B. Anal. Chem. 1996, 68, 4244–4253. (9) Lu, W.; Xu, X.; Cole, R. B. Anal. Chem. 1997, 69, 2478–2484. (10) Deng, H.; Van Berkel, G. J. Anal. Chem. 1999, 71, 4284–4293. (11) Deng, H.; Van Berkel, G. J.; Takano, H.; Gazda, D.; Porter, M. D. Anal. Chem. 2000, 72, 2641–2647. (12) Regino, M. C. S.; Brajter-Toth, A. Electroanalysis 1999, 11, 374–379. (13) Bo ¨kman, C. F.; Zettersten, C.; Sjo ¨berg, P. J. R.; Nyholm, L. Anal. Chem. 2004, 76, 2017–2024. (14) Blades, A. T.; Ikonomou, M. G.; Kebarle, P. Anal. Chem. 1991, 63, 2109– 2114. (15) Van Berkel, G. J.; McLuckey, S. A.; Glish, G. L. Anal. Chem. 1992, 64, 1586–1593. (16) Van Berkel, G. J.; Zhou, F. Anal. Chem. 1995, 67, 2916–2923. (17) Deng, H.; Van Berkel, G. J. Electroanalysis 1999, 11, 857–865. (18) Jurva, U.; Wikstrom, H. V.; Bruins, A. P. Rapid Commun. Mass Spectrom. 2000, 14, 529–533. (19) Jurva, U.; Wikstrom, H. V.; Weidolf, L.; Bruins, A. P. Rapid Commun. Mass Spectrom. 2003, 17, 800–810. (20) Roussel, C.; Dayon, L.; Lion, N.; Rohner, T. C.; Josserand, J.; Rossier, J. S.; Jensen, H.; Girault, H. H. J. Am. Soc. Mass Spectrom. 2004, 15, 1767– 1779. (21) Permentier, H. P.; Jurva, U.; Barroso, B.; Bruins, A. P. Rapid Commun. Mass Spectrom. 2003, 17, 1585–1592. (22) Permentier, H. P.; Bruins, A. P. J. Am. Soc. Mass Spectrom. 2004, 25, 1707– 1716. (23) Diehl, G.; Karst, U. Anal. Bioanal. Chem. 2002, 373, 390–398. (24) Karst, U. Angew. Chem., Int. Ed. 2004, 43, 2476–2478. 10.1021/ac901975j CCC: $40.75  2009 American Chemical Society Published on Web 10/29/2009

Ambient mass spectrometry25,26 has recently been introduced to provide direct ionization of analytes with little or no sample preparation. Desorption electrospray ionization (DESI)27 developed by Cooks and co-workers and direct analysis in real time (DART)28 developed by Cody and co-workers were the first two of this new family of technologies. It has been shown that DESI is of great value in the fast analysis of a variety of different analytes ranging from pharmaceuticals to tissue imaging.29-37 In DESI experiments, ionization of samples occurs via the interactions with charged microdroplets generated in a pneumatically assisted electrospray of an appropriate solvent. Recently, there have been several accounts reporting the unexpected oxidation of analytes observed during ionization by DESI.38-40 In addition to being used regularly for solid sample analysis from surfaces, DESI has been extended to allow the direct analysis of liquid samples.41-45 It has been found that liquid sample DESI can be used to analyze a wide range of nonvolatile compounds including high-mass proteins without sample preparation, enabling their fast chemical analysis.44 In the previously reported studies for the combination of EC with ESI,6,8 the EC system usually needed to be either floated or decoupled from the ESI source since ESI operates at a high voltage whereas electrochemical cells are operated at a low voltage. In our preliminary trial (i.e., the DESI-MS detection of perylene radical cation generated by electro-oxidation of perylene in a nonaqueous solution in a tubular cell),44 liquid sample DESI was shown to be useful for this coupling purpose. When EC was combined with DESI-MS, the coupling is simple as the electrolyzed sample solution flowing out of electrochemical cell can be (25) Cooks, R. G.; Ouyang, Z.; Takats, Z.; Wiseman, J. M. Science 2006, 311, 1566–1570. (26) Venter, A.; Nefliu, M.; Cooks, R. G. Trends Anal. Chem. 2008, 27, 284– 290. (27) Taka´ts, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Anal. Chem. 2004, 76, 4050–4058. (28) Cody, R. B.; Laramee, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 2297– 2302. (29) Wiseman, J. M.; Ifa, D. R.; Song, Q.; Cooks, R. G. Angew. Chem., Int. Ed. 2006, 45, 7188–7192. (30) Dixon, R. B.; Bereman, M. S.; Muddiman, D. C.; Hawkridge, A. M. J. Am. Soc. Mass Spectrom. 2007, 18, 1844–1847. (31) Bereman, M. S.; Williams, T. I.; Muddiman, D. C. Anal. Chem. 2007, 79, 8812–8815. (32) Van Berkel, G. J.; Ford, M. J.; Deibel, M. A. Anal. Chem. 2005, 77, 1207– 1215. (33) Yang, S.; Han, J.; Huan, Y.; Cui, Y.; Zhang, X.; Chen, H.; Gu, H. Anal. Chem. 2009, 81, 6070–6079. (34) Lane, A.; Nyadong, L.; Galhena, A. S.; Shearer, T. L.; Stout, E. P.; Parry, R. M.; Kwasnik, M.; Wang, M. D.; Hay, M. E.; Fernandez, F. M.; Kubanek, J. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 7314–7319. (35) Denes, J.; Katona, M.; Hosszu, A.; Czuczy, N.; Takats, Z. Anal. Chem. 2009, 81, 1669–1675. (36) Meetani, M. A.; Shin, Y.-S.; Zhang, S.; Mayer, R.; Basile, F. J. Mass Spectrom. 2007, 42, 1186–1193. (37) Pol, J.; Novak, P.; Volny, M.; Kruppa, G. H.; Kostiainen, R.; Lemr, K.; Havlicek, V. Eur. J. Mass Spectrom. 2008, 14, 391–399. (38) Pasilis, S. P.; Kertesz, V.; Van Berkel, G. J. Anal. Chem. 2008, 80, 1208– 1214. (39) Volny´, M.; Venter, A.; Smith, S. A.; Pazzi, M.; Cooks, R. G. Analyst 2008, 132, 525–531. (40) Benassi, M.; Wu, C.; Nefliu, M.; Ifa, D.; Volny, M.; Cooks, R. G. Int. J. Mass Spectrom. 2009, 280, 235–240. (41) Mulligan, C. C.; MacMillan, D. K.; Noll, R. J.; Cooks, R. G. Rapid Commun. Mass Spectrom. 2007, 21, 3729–3736. (42) Ma, X.; Zhao, M.; Lin, Z.; Zhang, S.; Yang, C.; Zhang, X. Anal. Chem. 2008, 80, 6131–6136. (43) Chipuk, J. E.; Brodbelt, J. S. J. Am. Soc. Mass Spectrom. 2008, 19, 1612– 1620. (44) Miao, Z.; Chen, H. J. Am. Soc. Mass Spectrom. 2009, 20, 10–19.

Figure 1. Scheme showing the apparatus for online coupling of a thin-layer electrochemical flow cell with DESI-MS. WE, working electrode; AE, auxiliary electrode; and RE, reference electrode.

directly desorbed and ionized by DESI-MS. In this study, we adopted an optimized configuration for liquid sample DESI (in comparison to the previous configuration,44 we simply removed the sample surface and directly directed the DESI spray to the capillary tip from which sample solution flowed out; this modification allows easy position adjustment and high sensitivity) and coupled a thin-layer electrochemical flow cell online with liquid sample DESI (Figure 1). Employing such an EC/DESI-MS coupling device, we investigated several electrochemical reactions involving biological molecules in aqueous solution, including dopamine and thiol oxidation, peptide and protein disulfide bond reduction, and online derivatization of peptides and proteins. In this article, we report the characteristics and some applications of the EC/DESI-MS coupling, based on those reactions examined. EXPERIMENTAL SECTION Chemicals. Dopamine hydrochloride, cysteine, glutathione disulfide (GSSG), peptide (sequence: NRCSQSCWN), insulin from bovine pancreas, and β-lactoglobulin A from bovine milk were purchased from Sigma-Aldrich (St. Louis, MO) and used without further purification. Glutathione (GSH) was purchased from TCI (Portland, OR) and used as received. HPLC-grade methanol was purchased from GFS Chemicals (Columbus, OH). The deionized water used for sample preparation was obtained using a Nanopure Diamond Barnstead purification system (Barnstead International, Dubuque, IA). Online EC/DESI-MS Apparatus. Figure 1 shows a homebuilt apparatus for coupling a commercial thin-layer electrochemical flow cell (BAS, West Lafayette, IN) with a hybrid triplequadrupole-linear ion trap mass spectrometer (Q-trap 2000, Applied Biosystems/MDS SCIEX, Concord, Canada) using liquid sample DESI. The thin-layer electrochemical flow cell consisted of a working electrode (WE) embedded in PEEK and separated from a stainless steel auxiliary electrode (AE) by a Teflon gasket (0.01 in. thick) and an Ag/AgCl (3 M NaCl) reference electrode (RE) contacting the sample solution through a small hole in the AE. The WE used was either a glassy carbon electrode (6 mm diameter, BAS, West Lafayette, IN) for oxidation reactions or a dual amalgam electrode (3 mm diameter) for reduction reactions. The amalgam electrode was prepared by depositing a thin-layer of mercury onto a dual gold electrode surface (BAS, West Lafayette, IN). A potentiostat (CV-27 model, BAS, West Lafayette, IN) was used to apply the potential to the cell to trigger Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

9717

Figure 2. (+)-DESI MS spectra acquired when dopamine (1.0 mM) in H2O/CH3OH (1:1 by volume) containing 1% acetic acid flowed through the thin-layer electrochemical cell with an applied potential of (a) 0.0 and (b) 1.5 V.

oxidization/reduction of analytes that flow through the electrochemical cell. Sample solutions were degassed by argon purging for 30 min to remove dissolved oxygen prior to the injection to the flow cell for electrolysis. The DESI spray probe was aimed at the mass spectrometer’s inlet orifice and kept 3-4 cm away from the orifice (Figure 1). The oxidized/reduced species flowed out of the thin-layer cell via a short piece of fused silica connection capillary (i.d. 0.05 or 0.1 mm, 3.8 cm long) and underwent interactions with the charge microdroplets from DESI spray for ionization. The capillary outlet was placed about 2-3 mm downstream from the DESI spray probe tip and kept in line with the sprayer tip and the mass spectrometer’s inlet orifice (as mentioned above, this is an improved configuration for liquid sample DESI in which the sample surface is no longer used). The spray solvent for DESI was H2O/CH3OH (1:1 by volume) containing 1% acetic acid and the spray solution injection rate was 5-10 µL/min. The flow rate for sample solutions passing through the electrochemical cell for electrolysis was 5 µL/min, unless specified otherwise. A high voltage of -4.5 or + 5 kV was applied to the spray probe, depending on the polarity of the DESI ion mode used. The mass spectrometer curtain gas (N2) was kept as 10-20 (manufacturer’s units), and the declustering potential was set at 10-50 V. Collision induced dissociation (CID) was carried out to provide ion structural information using enhanced product ion scan mode, and N2 was used as the collision gas. Data acquisition was performed using the Analyst software (version 1.4.2, Applied Biosystems/MDS SCIEX, Concord, Canada). Deconvolution of mass spectra was carried out using MagTran 1.03 software (Amgen Inc., Thousand Oaks, CA) written based on the ZScore algorithm.46 (45) Zhang, Y.; Chen, H. Int. J. Mass Spectrom. 2009, in press. (46) Zhang, Z.; Marshall, A. J. Am. Soc. Mass Spectrom. 1998, 9, 225–233.

9718

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

RESULTS AND DISCUSSION Electrochemical Oxidation Reactions. Dopamine, a classical compound often used for the characterization of liquid chromatography/electrochemistry systems (the oxidation of this compound is shown in the inset of Figure 2a), was first chosen to examine the performance of the EC/DESI-MS coupling apparatus (Figure 1). In the experiment, when a 1.0 mM solution of dopamine (MW 153) was allowed to flow through the thin-layer electrochemical cell without any potential applied and followed by online DESI ionization, the mass spectrum shows the protonated dopamine at m/z 154 (Figure 2a). Upon CID, it gives rise to fragment ions of 137, 119, and 91 by consecutive losses of NH3, H2O, and CO, confirming its structure. In Figure 2a, another ion of m/z 137 was also observed; this is due to the dissociation of the protonated dopamine (m/z 154) during the ion transmission inside the mass spectrometer (the enhanced mass scan mode was used for data acquisition, in which collision gas was used to help ion transmission in the Qtrap instrument used). While the electrochemical cell was turned on (1.5 V), the protonated ion (m/z 152) of the oxidized dopamine product, dopamine o-quinone (DQ), was detected (Figure 2b). The CID dissociation of m/z 152 produces mainly the fragment ion of m/z 123 by loss of CH2dNH, which also appears in Figure 2b. These results clearly show that the oxidized dopamine, a nonvolatile species, generated via electrolysis in the cell can be directly ionized and transferred from the solution phase into the gas phase for MS detection by DESI. The inset of Figure 2b further illustrates the extracted ion chromatogram of m/z 152 recorded, reflecting the good MS response toward the applied cell potential. The response time of the EC/DESI-MS apparatus, defined as the temporal interval from the time that the cell was turned on to the time that the product ion of m/z 152 was detected, was measured to be 3.6 s when the sample flow rate was 30 µL/min. The fast response of EC/DESIMS can be accounted for by the fact that the connection capillary

bridging the cell and DESI source is short and small (3.8 cm in length and 50 µm in diameter). In this study, we used a commercial thin-layer flow cell (known for easy cleaning and replacement17). However, one can imagine that such a flow cell could be further modified so that no connection conduit is necessary, which would further shorten the response time of EC/ DESI-MS coupling. One particular phenomenon that we noticed is that there is no m/z 152 in Figure 2a, indicating that the ionization of dopamine by DESI does not result in oxidation. This result is different from a previous study on the oxidation of dopamine using EC/ESIMS17 in which a small fraction of dopamine was shown to be oxidized in the ESI emitter. This feature of EC/DESI-MS is advantageous as background signal (when the cell is off) is undesirable. Recently, there have been several accounts reporting the unexpected oxidation observed from DESI analysis of solid samples on a surface.38-40 In our experiment using DESI to desorb and ionize dopamine in solution, the oxidation during ionization does not take place probably either due to the buffer effect of the solution solvent or because no high voltage was directly applied to the solution sample. In addition, as the DESI ionization event is separated from the electrolysis process in our experiment, one could use traditional solvent systems (e.g., water without the addition of organic solvents) for electrolysis. We tested the electro-oxidation of dopamine (1.0 mM) in H2O containing 10 mM NaCl as the electrolyte in the thin-layer electrochemical cell. At a cell potential of 1.4 V, the obtained (+)-DESI MS spectrum shows the formation of the protonated ion of the oxidized dopamine product (m/z 152) along with its fragment ion at m/z 123 (Figure 1Sa, Supporting Information). Also, presumably because of the use of sodium chloride NaCl as an electrolyte, the oxidation yield of 45%, defined as the combined abundance of the product ions of m/z 123 and 152 divided by the total abundance of the ions of m/z 123, 137, 152, and 154, is higher than 25% oxidation conversion yield in the experiment mentioned above using acetic acid as an electrolyte (Figure 2b). Furthermore, we tested the oxidation of dopamine using a porous electrode flow cell (model 5030, ESA Biosciences Inc., Chelmsford, MA), and it shows that the conversion yield can be improved to be 61% (Figure 1Sb, Supporting Information). Glutathione (GSH, MW 307 Da), a cysteine-containing tripeptide, was also examined by EC/DESI-MS. This compound is known to undergo reversible oxidation to produce glutathione disulfide (GSSG, MW 612 Da, Figure 3a). Figure 3b depicts the MS spectrum acquired when a GSH solution passed through the cell (no potential was applied) followed by DESI-MS analysis. One can see the protonated ion [GSH + H]+ at m/z 308 and the dimer ion [2GSH + H]+ at m/z 615. While a potential of 2.0 V was applied to the cell, the singly and doubly charged ions of the oxidized peptide GSSG arose at m/z 613 and 307, respectively (Figure 3c). The CID spectrum of m/z 613 shows the backbond cleavage of the disulfide-bond linked peptide via losses of water, glycine, and alanyl ketene, consistent with the ion structural assignment (Figure 2S, Supporting Information). This result shows the oxidative formation of a disulfide bond of GSSG from free thiol groups of GSH molecules occurred in the electrochemical

Figure 3. (a) Scheme showing the oxidative/reductive conversion between glutathione (GSH) and glutathione disulfide (GSSG); (+)-DESI MS spectra acquired when GSH (1.0 mM) in H2O/CH3OH (1:1 by volume) containing 1% acetic acid flowed though the thin-layer electrochemical cell with the applied potential of (b) 0.0 and (c) 2.0 V.

cell and the reaction products can be monitored using direct online DESI-MS analysis. Electrochemical Reduction Reactions: Disulfide Bond Cleavage. Cleavage of disulfide bonds is often in need for structural analysis of peptides and proteins because it provides a number of smaller peptides whose sequencing can aid the structure determination of the original peptide/protein.47 The traditional protocol to break disulfide bonds is chemical reduction using reagents like dithiothreitol (DTT). However, the elimination of byproducts is timeconsuming and troublesome. Recently, there have been tremendous amounts of effort on the disulfide bond cleavage based on newly discovered ion chemistry.48-53 An alternative way for eliminating disulfide bond without producing byproducts or involving chemical reagents is electrolytic reduction.54,55 In our study, EC/DESI-MS allows the direct electrochemical reduction of peptide and protein disulfide bonds followed with MS analysis of the reduction products. (47) Bilusich, D.; Bowie, J. H. Mass Spectrom. Rev. 2009, 28, 20–34. (48) Gunawardena, H. P.; O’Hair, R. A. J.; McLuckey, S. A. J. Proteome Res. 2006, 5, 2087–2092. (49) Chrisman, P. A.; McLuckey, S. A. J. Proteome Res. 2002, 1, 549–557. (50) Kim, H. I.; Beauchamp, J. L. J. Am. Soc. Mass Spectrom. 2009, 20, 157– 166. (51) Li, J.; Shefcheck, K.; Callahan, J.; Fenselau, C. Int. J. Mass Spectrom. 2008, 278, 109–113. (52) Xia, Y.; Cooks, R. G. J. Am. Chem. Soc. 2009, submitted. (53) Qiao, L.; Bi, H.; Busnel, J.-M.; Liu, B.; Girault, H. H. Chem. Commun. 2008, 47, 6357–6359. (54) Dohan, J. S.; Woodward, G. E. J. Biol. Chem. 1939, 129, 393–403. (55) Cayot, P.; Rosier, H.; Roullier, L.; Haertle, T.; Tainturier, G. Food Chem. 2002, 77, 309–315.

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

9719

Figure 4. (a) (+)-DESI MS spectra acquired when a solution of GSSG (0.1 mM) in H2O containing 1% acetic acid flowed through the thin-layer electrochemical cell with an applied potential of 0.0 V (shown in the discrete line) and -1.6 V (shown in the solid line); (b) (-)-DESI MS spectra acquired when a solution of GSSG (0.1 mM) in H2O containing 1% acetic acid flowed through the thin-layer electrochemical cell with an applied potential of 0.0 V (shown in the discrete line) and -1.6 V (shown in the solid line). The discrete and solid lines were superimposed for comparison after their intensities were normalized.

Figure 5. (+)-DESI MS spectra acquired when a solution of insulin (0.1 mM) in H2O/CH3OH (1:1 by volume) containing 1% acetic acid flowed through the thin-layer electrochemical cell with an applied potential of (a) 0.0 and (b) -1.5 V. The inset in part a shows the structure of intact insulin which contains three disulfide bonds. 9720

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

Glutathione disulfide (GSSG), an oxidized form of the cysteinecontaining tripeptide glutathione (GSH) as mentioned above, was tested using the EC/DESI-MS for this purpose. In our experiments, we prepared an amalgam working electrode for use by depositing a thin-layer of mercury onto a gold electrode surface. First we tried to use the positive (+)-DESI ion mode (i.e., applied +5 kV to the DESI spray probe and monitored the resulting positive ions) and acquired mass spectra when a solution of glutathione disulfide (GSSG) (0.1 mM) in H2O containing 1% acetic acid flowed through the thinlayer cell (Figure 4a, note that this is another example of electrolyzing sample in water without addition of organic solvents). When no potential was applied to the cell, the doubly charged ion [GSSG + 2H]2+ was detected, which appears at m/z 307.0 along with its two major isotope peaks at m/z 307.5 and 308.0, respectively (as shown in the discrete line, Figure 4a). The first isotope peak is contributed from C13 while the second one is from S.34 When a -1.6 V potential was applied to the working electrode of the cell, the abundance of m/z 307.0 and 307.5 decreased, indicating the consumption of GSSG during electrolysis (shown in the solid line, Figure 4a). However, the signal of m/z 308.0 had a marginal increase. Because of the fact that the singly charged ion of glutathione GSH (the reduced product of GSSG from electrolysis), [GSH + H]+, overlaps with the second isotope peak of the doubly charged ion [GSSG + 2H]2+ at m/z 308.0, it is very likely that the increase in the intensity of m/z 308.0 from the formation of GSH is partially offset by the decrease from the depletion of GSSG by reduction. In other words, as a result of the peak overlap, it is hard to explicitly observe the formation of the reduced peptide product GSH. Because GSH carries two carboxylic acid groups (structure shown in Figure 3a), it is easy to form negative ions. Also, in our EC/DESI-MS coupling, there is no restriction to the polarity of DESIMS regardless of the polarity of the working electrode used since the DESI probe is physically separated from the cell. Thus we attempted to use the negative DESI-MS ion mode to detect the reduced product. Figure 4b (the discrete line) shows the (-)-DESI MS spectrum recorded when the cell was off, in which there is no detectable signal at m/z 306, corresponding to deprotonated GSH ion [GSH - H]-. Again, the absence of m/z 306 suggests that no reduction occurs during the ionization of GSSG by DESI, either, in agreement with our previous observed inhibition of reduction in the negative DESI ion mode.44 When the cell was on, the signal of m/z 306 arose (shown in the solid line, Figure 4b). Upon CID, this ion dissociates via consecutive losses of H2S, H2O, CO2, glycine, and gylcyl ethylene (Figure 3Sa, Supporting Information), exactly matching the CID of m/z 306 [GSH - H]generated by ESI of authentic compound GSH (Figure 3Sb, Supporting Information) and confirming the successful reduction of GSSG into GSH in our experiment. Note that there is one adjacent ion of m/z 297 from the background (Figure 4b), which remained almost unchanged during electrolysis. Electrolytic reduction of disulfide bonds of proteins was also explored using the EC/DESI-MS apparatus. Bovine pancreas insulin (structure is shown in the inset of Figure 5a, MW 5733.6 Da) is known to have two peptide chains linked by two disulfide bonds, and the A chain of the insulin has an additional disulfide bond. Figure 5a illustrates the (+)-DESI MS spectrum acquired when a solution of insulin (0.1 mM) in H2O/CH3OH (1:1 by

Figure 6. (a) (+)-DESI MS spectrum acquired when dopamine (0.5 mM) in the presence of cysteine (0.5 mM) in H2O/CH3OH (1:1 by volume) containing 1% acetic acid flowed through the thin-layer electrochemical cell with an applied potential of 1.0 V; the product of online derivatization of cysteine is observed at m/z 273; (b) (+)-DESI MS spectrum acquired when dopamine (0.5 mM) in the presence of a peptide (amino acid sequence, NRCSQSCWN; MW 1154.2 Da) (50 µM) in H2O/CH3OH (1:1 by volume) containing 1% acetic acid flowed through the thin-layer electrochemical cell with an applied potential of 1.2 V; the tagged peptide product ions are observed; M in the figure represents the peptide; (+)-DESI MS spectra acquired when a solution of dopamine (0.5 mM) in the presence of β-lactoglobulin A (20 µM) in H2O/CH3OH (1:1 by volume) containing 1% acetic acid flowed through the thin-layer electrochemical cell with an applied potential of (c) 0.0 and (d) 1.2 V. The tagged protein ions are clearly observed and indicated with +11′ to +17′ ions. The peaks labeled with asterisks might be due to the noncovalent binding of two molecules of unoxidized dopamine by intact protein ions as the mass shift is 307.1 (see the deconvoluted spectrum in the inset), roughly equivalent to two molecules of dopamine (MW 153).

volume) containing 1% acetic acid flowed through the thin-layer cell without an applied potential to the cell. The multiply charged protein ion [insulin + 6H]6+ at m/z 956.6 was detected (+4 and +5 insulin ions were also detected in the higher mass range which are not shown in Figure 5a). When a -1.5 V potential was applied to the amalgam working electrode, ions of m/z 681.4 and 851.0 arose, corresponding to the +5 and +4 ions of the separated insulin B chain (MW 3399.9 Da). Also, an ion of m/z 781.0 corresponding to +3 ions of the fully reduced and separated insulin A chain (MW 2339.7 Da) was observed. The relative intensity of the A chain ions is lower than those of the B chain ions. It is probably because the A chain has several acidic amino acid residues and lacks basic residues such as arginines, histidines, and lysines so that it has relatively low proton affinity. The simultaneous detection of both fully reduced A and B chains of insulin shows that all three disulfide bonds of insulin underwent cleavage via electrolytic reduction. The successful electrolytic reduction of disulfide bonds of both peptides and proteins followed with online MS detection as demonstrated in this study using EC/DESI-MS would be invaluable in MS-based proteomics research. To the best of our knowledge, this is also one of few reduction examples shown in the online coupling of electrochemistry with mass spectrometry.56

Online Derivatization Reactions Using Electrogenerated Species. Online mass tagging is useful for chemical analysis of mixtures because such a derivatization could enhance the sensitivity, selectivity, or both for the analyte detection. For example, online tagging of cysteine has been explored using inherent electrochemistry of ESI20 or EC/ESI-MS coupled devices.17 In this study, we also investigated the tagging applications using EC/ DESI-MS. First we tested the amino acid cysteine. A mixture of dopamine (0.5 mM) and cysteine (0.5 mM) in H2O/CH3OH (1:1 by volume) containing 1% acetic acid was injected into the thin-layer electrochemical cell for oxidation with the applied cell potential of 1.0 V and then followed with DESI analysis. As shown in the resulting MS spectrum (Figure 6a), the derivatized cysteine product ions, with a mass increment of 151 Da in comparison with the underivatized cysteine ion of m/z 122, appears at m/z 273. This product stems from the nucleophilic addition of DQ (MW 151 Da) generated from electro-oxidation of dopamine by cysteine (the suggested chemical equation is shown in the inset of Figure 6a). The peak of m/z 152 [DQ + H]+ is missing in Figure 6a, indicating that all of the electro-generated DQ has been reacted with cysteine. We further tested a peptide with amino acid sequence of NRCSQSCWN (MW 1154.2 Da), which contains two free cysteine Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

9721

residues. Figure 6b illustrates the (+)-DESI MS spectrum obtained when dopamine (0.5 mM) in the presence of the peptide (50 µM) in H2O/CH3OH (1:1 by volume) containing 1% acetic acid flowed through the cell with an applied potential of 1.2 V. Both doubly and triply charged ions of the tagged peptide product are observed, as labeled in the spectrum. Also, single and double additions of DQ to cysteines of the peptide occurred, as the peptide used contains two free cysteines. For instance, the ion of m/z 486.3 [M + 2DQ + 3H]3+ corresponds to the triply charged product produced from the double additions of DQ onto the peptide. These results are in good agreement with the previous report of online tagging using inherent electrochemistry of ESI.20 In the case of proteins, we examined β-lactoglobulin A, a protein containing only one free cysteine residue. In this experiment, a solution of dopamine (0.5 mM) doped with β-lactoglobulin A (20 µM, MW 18.4 kDa) was injected into the thin-layer cell (potential was off). Figure 6c displays the obtained (+)-DESI MS spectrum showing the intact multiply charged ions +11 to +18 of the protein β-lactoglobulin A. In this EC/DESI-MS experiment, the ionization of proteins by DESI benefits from the fact that the protein existed in solution when it flowed out of the cell, since desorption and ionization of proteins, especially those with high masses, by DESI from solution is relatively easier than from a solid sample.44 When the cell potential was applied, the oxidation of dopamine occurred and the oxidized dopamine reacted with the protein giving rise to the derivatized protein, as evidenced by the newly appeared peaks ranging from +11′ to +17′ (Figure 6d) and the deconvoluted spectrum shown in the figure inset. In the deconvoluted spectrum, the mass difference between untagged and tagged protein ions is 151.8 Da, corresponding to the addition of one molecule of DQ to the sole cysteine site of the protein (the MW of DQ is 151 Da and the small mass error of 0.8 Da might be due to the low mass resolution of our Qtrap instrument). Besides the tagged protein ions, there are additional satellite peaks (labeled with asterisks) associated with the protein ions +12, +13, and +14. It is likely that these peaks arise from the noncovalent binding of two molecules of unoxidized dopamine by intact protein ions as the mass shift is 307.1 (calculated using the deconvoluted mass spectrum), roughly equivalent to two molecules of dopamine (MW 153). The capability of online tagging shown in the EC/ DESI-MS coupling might find applications in pinpointing or even counting the cysteine sites of peptides and proteins. Also, electrochemical preparation of the tagging reagent online provides the ability to quickly turn on and off the tagging reaction by electronic means.57

9722

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

CONCLUSIONS In summary, the online coupling of EC/DESI-MS is a versatile platform to explore various electrochemical reactions of biological molecules. With the benefit from the DESI strength in direct sampling, the coupling of EC with MS is simple and has advantages of fast response and freedom to choose favorable ionization modes and solvent systems. Also, the absence of background signal possibly resulting from oxidation/reduction in DESI ionization itself is advantageous. However, the electrochemical conversion yields shown in the work are not high, in general, and future possible solutions include the usage of porous working electrodes or small amount of traditional strong electrolytes such as NaCl (as we demonstrated in this work) or designing new cells that allow direct desorption from working electrodes to avoid the reverse reactions occurring on auxiliary electrodes. Nevertheless, the demonstrated capability of electrochemical oxidation formation and reductive cleavage of disulfide bonds in EC/DESI-MS is of high value in future structural analysis of both peptides and proteins, as the disulfide bond is an important type of posttranslational modification of proteins. Likewise, online residue tagging using electrogenerated reactive species is important in probing specific residue sites of proteins. These results suggest that the online EC/DESI-MS coupling has potential in finding novel analytical applications. ACKNOWLEDGMENT The authors are grateful for helpful discussions with Dr. Jonathon O. Howell (BAS, West Lafayette, IN) and Professor Evan R. Williams (UC Berkeley) and for help from Zhixin Miao. J.L. thanks SINOPEC Shanghai Research Institute of Petrochemical Technology for the financial support to visit Ohio University. This work was supported by NSF (Grant CHE-0911160) and Ohio University (Grant SU1006172). SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review September 2, 2009. Accepted October 20, 2009. AC901975J (56) Johnson, K. A.; Shira, B. A.; Anderson, J. L.; Amster, I. J. Anal. Chem. 2001, 73, 803–808. (57) Van Berkel, G. J.; Kertesz, V. Rapid Commun. Mass Spectrom. 2009, 23, 1380–1386.