Electrochemically Induced pH Changes Resulting in Protein Unfolding

Anal. Chem. 2001, 73, 4836-4844

Electrochemically Induced pH Changes Resulting in Protein Unfolding in the Ion Source of an Electrospray Mass Spectrometer Lars Konermann,* Eloi A. Silva,† and Olusola F. Sogbein

Department of Chemistry, The University of Western Ontario, London, ON, N6A 5B7, Canada

The operation of an electrospray ion source in the positive ion mode involves charge-balancing oxidation reactions at the liquid/metal interface of the sprayer capillary. One of these reactions is the electrolytic oxidation of water. The protons generated in this process acidify the analyte solution within the electrospray capillary. This work explores the effects of this acidification on the electrospray ionization (ESI) mass spectrum of the protein cytochrome c (cyt c). In aqueous solution containing 40% propanol, cyt c unfolds around pH 5.6. Mass spectra recorded under these conditions, using a simple ESI series circuit, display a bimodal charge-state distribution that reflects an equilibrium mixture of folded and unfolded protein in solution. These spectra are not strongly affected by electrochemical acidification. An “external loop” is added to the ESI circuit when the metal needle of the sample injection syringe is connected to ground. The resulting circuit represents two coupled electrolytic cells that share the ESI capillary as a common anode. Under these conditions, the rate of charge-balancing oxidation reactions is dramatically increased because the ion source has to supply electrons for both, the external circuit and the ESI circuit. The analytical implications of this effect are briefly discussed. Mass spectra of cyt c recorded with the syringe needle grounded are shifted to higher charge states, indicating that electrochemical acidification has caused the protein to unfold in the ion source. The acidification can be suppressed by increasing the flow rate and lowering the electrolyte concentration of the solution and by using an electrolyte that acts as redox buffer. The observed acidification is similar for sprayer capillaries made of platinum and stainless steel. Removal of the protective oxide layer on the stainless steel surface results in effective redox buffering for a few minutes. Electrospray ionization (ESI) mass spectrometry (MS) is a versatile and highly sensitive method for the detection, quantitation, and structural analysis of a wide variety of analytes, including large biomolecules such as nucleic acids and proteins. 1-5 During * Corresponding author: (e-mail) [email protected]; (www) http:// publish.uwo.ca/∼konerman/. † Permanent address: Department of Chemistry, Federal University of Espirito Santo, Av. Fernando Ferrari, Cep. 29060-900, Vitoria-ES, Brazil. (1) Bruins, A. P.; Covey, T. R.; Henion, J. D. Anal. Chem. 1987, 59, 26422646.

4836 Analytical Chemistry, Vol. 73, No. 20, October 15, 2001

ESI, intact gas-phase ions are generated from analyte molecules in solution. In the commonly used positive ion mode, this process normally involves analyte protonation. It has been recognized that an ESI source can be described as a controlled-current electrolytic cell.6-8 Analyte solution is pumped through a metal capillary to which a positive electric potential (typically around +5000 V) is applied. Electrophoretic charge separation leads to the emission of positively charged solvent droplets from a Taylor cone at the tip of the ESI capillary. These droplets rapidly shrink due to solvent evaporation, thus increasing the charge density on the droplet surface and eventually leading to the formation of offspring droplets via jet fission. Analyte gas-phase ions are generated, either as charged residues from nanometer-sized offspring droplets or by ion evaporation.8-11 These ions are sampled by an atmospheric pressure-to-vacuum interface and transferred to the mass analyzer. To maintain charge balance, the generation of positively charged solvent droplets from an electrically neutral solution has to be accompanied by electrolytic oxidation reactions at the metal/liquid interface of the ESI capillary. Electrons produced in these reactions flow to the high-voltage power supply. They are passed on to the mass spectrometer to neutralize (i.e., reduce) the positive charges produced by the ESI source, thus completing an electrical series circuit.6-8,12-16 Under normal operating conditions, the “lowvoltage side” of the power supply is connected to ground (i.e., a potential of 0 V), as indicated in Figure 1. (2) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science 1989, 246, 64-71. (3) Smith, R. D.; Loo, J. A.; Edmonds, C. G.; Barinaga, C., J.; Udseth, H. R. Anal. Chem. 1990, 62, 882-899. (4) Cole, R. B. Electrospray Ionization Mass Spectrometry; John Wiley & Sons: New York, 1997. (5) Mann, M.; Wilm, M. Trends Biochem. Sci. 1995, 20, 219-224. (6) Blades, A. T.; Ikonomou, M. G.; Kebarle, P. Anal. Chem. 1991, 63, 21092114. (7) Van Berkel, G. J.; Zhou, F. Anal. Chem. 1995, 67, 2916-1923. (8) Kebarle, P.; Ho, Y. In Electrospray Ionization Mass Spectrometry; Cole, R. B., Ed.; John Wiley & Sons: New York, 1997; pp 3-63. (9) Kebarle, P.; Peschke, M. Anal. Chim. Acta 2000, 406, 11-35. (10) Juraschek, R.; Dulcks, T.; Karas, M. J. Am. Soc. Mass Spectrom. 1999, 10, 300-308. (11) Gamero-Castano, M.; de la Mora, J. F. Anal. Chem. 2000, 72, 1426-1429. (12) Kebarle, P.; Tang, L. Anal. Chem. 1993, 65, 972-986. (13) Van Berkel, G. J. In Electrospray Ionization Mass Spectrometry; Cole, R. B., Ed.; John Wiley & Sons: New York, 1997,7; pp 65-105. (14) Van Berkel, G. J.; Giles, G. E.; Bullock, J. S.; Gray, L. J. Anal. Chem. 1999, 71, 5288-5296. (15) Van Berkel, G. J.; De La Mora, J. F.; Enke, C. G.; Cole, R. B.; MartinezSanchez, M.; Fenn, J. B. J. Mass Spectrom. 2000, 35, 939-952. (16) Jackson, G. S.; Enke, C. G. Anal. Chem. 1999, 71, 3777-3784. 10.1021/ac010545r CCC: $20.00

© 2001 American Chemical Society Published on Web 09/11/2001

Figure 1. Schematic diagram of the experimental setup used for this study. Arrows indicate the direction of electron flow: itot, total current; iesi, electrospray current; iext, external current. For details, see text.

Charge-balancing oxidation reactions at the metal/liquid interface of the ESI capillary are of considerable interest from an analytical point of view because they change the chemical composition of the analyte solution. In general, the reaction with the lowest electrochemical potential will provide most of the electrons needed to compensate the positive charges emitted from the capillary tip. These reactions can involve oxidation of the capillary material (usually stainless steel), the analyte, and the solvent. Blades et al.6 showed that Zn2+ is released from an ESI capillary made of Zn when spraying solutions of various salts in methanol. Zn is a metal that is very easily oxidized (Zn f Zn2+ + 2e-, E0 ) -0.76 V). Similarly, stainless steel capillaries were found to release Fe2+ when methanol6 or acetonitrile solutions were sprayed17 (Fe f Fe2+ + 2e-, E0 ) -0.41 V). When the electrochemical potential of the analyte is lower than that of the capillary material, analyte oxidation can be the dominant chargebalancing reaction. This effect leads to the formation of radical cations (A•+) from neutral compounds and thus allows the ESI MS analysis of chemical species that cannot be ionized by protonation.18 This “electrochemical ionization” has been extensively utilized for the analysis of a variety of compounds.7,13,19-21 Electrolytic oxidation during ESI was also observed for a number of iron-containing metalloproteins.22-24 Of major interest for the current study is the electrolytic oxidation of water

2H2O f 4H+ + 4e- + O2

E0 ) +1.23 V

(1)

In the absence of easily oxidizable solutes, this is known to be the major charge-balancing reaction when aqueous solutions are sprayed from an ESI capillary made of platinum (Pt f Pt2+ + (17) Van Berkel, G. J. J. Anal. At. Spectrom. 1998, 13, 603-607. (18) Van Berkel, G. J.; McLuckey, S. A.; Glish, G. L. Anal. Chem. 1992, 64, 1586-1593. (19) Xu, X.; Nolan, S. P.; Cole, R. B. Anal. Chem. 1994, 66, 119-125. (20) McCarley, T. D.; Lufaso, M. W.; Curtin, L. S.; McCarley, R. L. J. Phys. Chem. B 1998, 102, 10078-10086. (21) Van Berkel, G. J.; Quirke, J. M. E.; Tigani, R. A.; Dilley, A. S.; Covey, T. R. Anal. Chem. 1998, 70, 1544-1554.

2e-, E° ≈ +1.2 V) or other oxidation-resistant materials.25 For systems where eq 1 represents the only charge-balancing reaction, the concentration of electrolytically produced protons, [H+]elec, can be calculated from Faraday’s law25

[H+ ]elec )

itot nFvf

(2)

where itot represents the electron current drawn from the ESI capillary, n ) 1 is the number of electrons produced per proton, F ) 96 485 C mol-1 is the Faraday constant, and vf is the flow rate of the solution that is pumped through the ESI capillary. Electrolytic water oxidation will cause a change in pH that is given by

∆pHmax ) - log([H + ]initial + [H + ]elec) - pHinitial (3) where [H+]initial and pHinitial are the proton concentration and the pH, respectively, of the solution before it enters the ESI capillary. The subscript “max” indicates that the actual pH change will often be smaller than predicted by eq 3 because (i) electrolytic oxidation of water may not be the only charge-balancing reaction and (ii) most solutions will have a certain pH buffering capacity. The current in a simple electrospray circuit does not depend very strongly on the solution flow rate as it is roughly proportional to vf0.5.8,26,27 Electrochemical acidification will therefore become more pronounced when the flow rate is decreased.25 Furthermore, (22) McLafferty, F. W.; Guan, Z.; Haupts, U.; Wood, T. D.; Kelleher, N. L. J. Am. Chem. Soc. 1998, 120, 4732-4740. (23) He, F.; Hendricksen, C. L.; Marshall, A. G. J. Am. Soc. Mass Spectrom. 2000, 11, 120-126. (24) Johnson, K. A.; Shira, B. A.; Anderson, J. L.; Amster, I. J. Anal. Chem. 2001, 73, 803-808. (25) Van Berkel, G. J.; Zhou, F.; Aronson, J. T. Int. J. Mass Spectrom. Ion Processes 1997, 162, 55-67. (26) Tang, K.; Lin, Y.; Matson, D. W.; Kim, T.; Smith, R. D. Anal. Chem. 2001, 73, 1658-1663. (27) Fernadez de la Mora, J.; Loscertales, I. G. J. Fluid Mech. 1994, 260, 155184.

Analytical Chemistry, Vol. 73, No. 20, October 15, 2001

4837

the largest pH changes are expected for near-neutral solutions. The occurrence of an electrochemical acidification during ESI was first demonstrated by Van Berkel and co-workers25 for aqueous solutions of various pH indicator dyes. By using an ESI source that was coupled to a spectrophotometer system, a pH reduction of up to 4 units was observed. The magnitude of the experimentally observed acidification was in good agreement with the predicted pH changes based on eq 3, indicating that water oxidation was the major charge-balancing reaction under the conditions of that study. Besides being a powerful analytical tool, ESI MS is also a wellestablished technique for monitoring protein folding and unfolding transitions in solution. The charge-state distribution generated during ESI is a sensitive probe for conformational changes of proteins. In general, an unfolded protein in solution will lead to the formation of higher charge states than the same protein in a tightly folded conformation.28-35 In the transition region, i.e., under conditions where both the folded and the unfolded conformation of the protein are populated, ESI mass spectra are characterized by a bimodal charge-state distribution.31,32,34 In contrast to other commonly used techniques for studying protein folding, ESI MS has the unique capability of monitoring noncovalent interactions in solution by direct observation of the corresponding gas-phase ion complexes.36 Therefore, ESI MS not only provides information on the protein conformation but also allows monitoring the loss or binding of protein ligands, metal ions, and changes of quaternary protein interactions during a folding or unfolding reaction.37-43 Additional information can be gained by coupling ESI MS on-line with hydrogen/deuterium exchange techniques.30,31,44-47 Acid is the most common denaturant for ESI MSbased protein-folding studies. At near-neutral pH, most proteins adopt a tightly folded conformation. The addition of acid to the solution will destabilize this native state and induce a transition to an unfolded and largely disordered conformation. (28) Chowdhury, S. K.; Katta, V.; Chait, B. T. J. Am. Chem. Soc. 1990, 112, 9012-9013. (29) Loo, J. A.; Edmonds, C. G.; Udseh, H. R.; Smith, R. D. Anal. Chem. 1990, 62, 693-698. (30) Katta, V.; Chait, B. T. J. Am. Chem. Soc. 1993, 115, 6317-6321. (31) Wagner, D. S.; Anderegg, R. J. Anal. Chem. 1994, 66, 706-711. (32) Konermann, L.; Douglas, D. J. Biochemistry 1997, 36, 12296-12302. (33) Konermann, L.; Collings, B. A.; Douglas, D. J. Biochemistry 1997, 36, 55545559. (34) Konermann, L.; Douglas, D. J. Rapid Commun. Mass Spectrom. 1998, 12, 435-442. (35) Pan, X. M.; Sheng, X. R.; Zhou, J. M. FEBS Lett. 1997, 402, 25-27. (36) Loo, J. A. Mass Spectrom. Rev. 1997, 16, 1-23. (37) Vis, H.; Heinemann, U.; Dobson, C. M.; Robinson, C. V. J. Am. Chem. Soc. 1998, 120, 6427-6428. (38) Sogbein, O. O.; Simmons, D. A.; Konermann, L. J. Am. Soc. Mass Spectrom. 2000, 11, 312-319. (39) Katta, V.; Chait, B. T. J. Am. Chem. Soc. 1991, 113, 8534-8535. (40) Veenstra, T. D.; Johnson, K. L.; Tomlinson, A. J.; Naylor, S.; Kumar, R. Biochemistry 1997, 36, 3535-3542. (41) Veenstra, T. D.; Johnson, K. L.; Tomlinson, A. J.; Craig, T. A.; Kumar, R.; Naylor, S. J. Am. Soc. Mass Spectrom. 1998, 9, 8-14. (42) Nemirovskiy, O. V.; Ramanathan, R.; Gross, M. L. J. Am. Soc. Mass Spectrom. 1997, 8, 809-812. (43) Lee, V. W. S.; Chen, Y.-L.; Konermann, L. Anal. Chem. 1999, 71, 41544159. (44) Wang, F.; Tang, X. Biochemistry 1996, 35, 4069-4078. (45) Woodward, C. J. Am. Soc. Mass Spectrom. 1999, 10, 672-674. (46) Babu, K. R.; Douglas, D. J. Biochemistry 2000, 39, 14702-14710. (47) Babu, K. R.; Moradian, A.; Douglas, D. J. J. Am. Soc. Mass Spectrom. 2001, 12, 317-328.

4838


Obviously, the occurrence of an electrochemical acidification during the spraying process is an important issue for proteinfolding studies by ESI MS. Keeping in mind the magnitude of the pH changes observed in ref 25 (up to 4 pH units), it seems likely that this effect could cause a protein to unfold in the ESI capillary, even when the pH of the solution entering the ion source is well below the transition midpoint for unfolding. This should lead to a disparity between ESI MS data and results obtained by methods where the protein structure is probed in bulk solution (e.g., optical techniques or NMR spectroscopy). Little is known about this issue and usually electrochemical effects are not discussed in ESI MS-based protein-folding studies. It has been shown that electrochemical acidification can enhance the signal intensity of protein mass spectra.25 However, it remains to be determined under which conditions this effect can cause the protein conformation, as probed by the ESI charge-state distribution, to change. In this work, we study the effects of various experimental parameters on the ESI mass spectrum of the protein cytochrome c (cyt c). Our data indicate that, under certain conditions, electrochemical processes can indeed cause a significant degree of protein unfolding in the ESI capillary. However, through the appropriate choice of experimental parameters, the effects of electrolytic pH changes in the ESI capillary can be largely eliminated. EXPERIMENTAL SECTION Chemicals. Horse heart cyt c (MW 12 360 2) was obtained from Sigma (St. Louis, MO). Glacial acetic acid and hydrochloric acid were obtained from BDH (Toronto, ON, Canada); 1-propanol and HPLC grade acetonitrile were supplied by Caledon (Georgetown, ON, Canada). All chemicals were used without further purification. Unless noted otherwise, the pH of all solutions was adjusted by the addition of acetic acid. pH measurements were carried out on an accumet AB15 pH meter (Fisher Scientific, Nepean, ON, Canada). Fluorescence measurements were performed on a Biologic (Claix, France) SFM-4S spectrometer. The protein solutions used in these experiments were identical to those used for ESI MS. The samples ware excited at 280 nm, and the total fluorescence intensity (λem g 325 nm) was monitored by a photomultiplier that had a cut-on filter mounted in front of it. Electrospray Mass Spectrometry. Ions were generated by pneumatically assisted ESI (“ion spray”)1,48 in the positive ion mode at an ion spray voltage of +5600 V. Purified air was used as nebulizer gas. The ESI capillaries (“sprayer capillaries”) used were made of platinum, type 304 stainless steel (both supplied by Hamilton, Reno, NE) or type 316 stainless steel (Small Parts Inc., Miami Lakes, FL). No differences were observed between types 304 and 316 steel; therefore, both will be referred to as “stainless steel” in the following text. Capillary dimensions were as follows: overall length 6 cm, i.d. 0.21 mm, and o.d. 0.41 mm. Protein solution was delivered to the mass spectrometer by a 1-mL glass syringe (SGE, Austin, TX) that was equipped with a type 304 stainless steel syringe needle. The syringe was mounted on a syringe pump (Harvard Apparatus, model 22, Saint Laurent, PQ, (48) Ikonomou, M. G.; Blades, A. T.; Kebarle, P. Anal. Chem. 1991, 63, 19891998.

Canada) and solution was pumped through a 75-cm-long fusedsilica capillary (TSP075150, Polymicro Technologies, Phoenix, AZ, i.d. 75 µm, o.d. 150 µm). One end of the fused-silica capillary was connected to the syringe needle by a PEEK connector (Upchurch Scientific, Oak Harbor, WA), the other end was located inside the ESI capillary. Unless stated otherwise, the fused-silica capillary was positioned flush with the ESI capillary (as indicated in Figure 1, this arrangement resulted in the best signal intensity and stability). The electric current was measured by a Fluke model 8060 A ammeter. This instrument is battery-powered and was operated well isolated from ground. These two factors are necessary for measuring electrical currents on the high-voltage side of the power supply. Mass spectra were recorded on a single quadrupole instrument (model Toby, Sciex, Concord, ON, Canada). The ion sampling interface of this instrument is essentially the same as that described in ref 49. The spectra shown in this work represent an average of at least six scans recorded with a dwell time of 10 ms per data point. RESULTS AND DISCUSSION Loop Currents in the Electrospray Circuit. A schematic diagram of the experimental setup used for this work is shown in Figure 1. Studies on the electrochemical aspects of the ESI process usually focus on the series circuit involving the high-voltage power supply, the ESI capillary, the charged droplets and ions that are emitted from the sprayer tip, and the mass spectrometer as counter electrode. The current in this “ESI loop circuit”, iesi, is symbolized by the solid arrows in Figure 1 (arrows are in the direction of electron flow). For experiments that involve continuous sample injection by a syringe pump, it is common practice to connect the stainless steel syringe needle to ground. The use of a grounding cable (dotted line in Figure 1) is strongly recommended by mass spectrometer manufacturers (see, e.g., Sciex API user manuals). It keeps the electric potential of the exposed syringe needle at 0 V and thus prevents the user from accidental contact with a hazardous high voltage. Figure 1 reveals that this procedure necessarily results in an “external loop circuit” with a current iext. The electron flow in the external loop circuit is indicated by the dotted arrows in Figure 1. Both loop circuits in this system share the ESI capillary, the high-voltage power supply, and the connections between these elements. The total electron current itot drawn from the ESI capillary is given by

itot ) iesi + iext

(4)

The magnitude of iext depends on the conductance G ) 1/R of the connection between the syringe needle and the ESI capillary. According to Ohm’s law, iext is given by

iext ) UG

(5)

where U is the voltage provided by the power supply. A nonzero current iext requires the presence of electrolytes in the analyte solution. When only one type of electrolyte is present, the conductance is given by (49) Hunter, C. H.; Mauk, A. G.; Douglas, D. J. Biochemistry 1997, 36, 10181025.

G ) κA/l

(6)

where l and A are the length and the cross-sectional area, respectively, of the capillary connection between the syringe needle and the ESI source. The magnitude of iext will therefore depend on the dimensions of the capillary used. Equation 6 is a simplification, since it is assumed that all other components of the external circuit have a zero resistance. Polarization effects at the metal/liquid interfaces are also neglected. The conductivity κ is given by

κ ) cΛm

(7)

where c is the concentration of the electrolyte and Λm is the molar conductivity.50 The magnitude of iext will therefore also depend on the type of electrolyte (i.e., whether it is fully or only partly ionized in solution) and on the electrolyte concentration. Measuring itot is a simple matter by operating an ammeter on the high-voltage side of the power supply. Most previous studies have used an arrangement where an ammeter was located between ground and the mass spectrometer.6-8,12,13,15,51 The current measured in those studies corresponds to iesi. It has the same magnitude as the current of positive charges that are emitted from the sprayer tip. The occurrence of a current in the external loop goes undetected in such an arrangement. Probably this is the reason why the occurrence of an external loop current has received little attention in the previous literature. It is noted that when the syringe needle is not grounded, iext is equal to zero and itot ) iesi. The effects of various parameters on the electrical currents in the system were explored by electrospraying aqueous solutions of 10 µM cyt c with 40% v/v propanol at pH 5.6. At a flow rate of 0.25 µL/min, a current itot of ∼40 nA was measured when the syringe needle was grounded. The contribution of iesi, measured by disconnecting the syringe needle from ground was ∼5 nA. When the exit of the ESI capillary was blocked so that the spray could not reach the mass spectrometer, the external current iext was determined to be ∼40 nA. These currents could be increased dramatically by adding a strong electrolyte (1 mM KNO3) to the solution. Under otherwise identical conditions, the values for itot and iext were found to be 300 and 230 nA, respectively, with iesi ∼60 nA. The measured iext in all these experiments did not depend on the solvent flow rate vf, whereas iesi showed the expected proportionality to vf0.5 8,26,27 (data not shown). The experimental error of the measured currents due to signal fluctuations was on the order of (15%. These data confirm that, within experimental error, itot is indeed the sum of the two currents iext and iesi (eq 4). This seemingly trivial result indicates that the two underlying processes are largely independent. Like the ESI loop, the external loop represents an electrolytic cell. A current in the external loop implies the occurrence of redox reactions at the metal/liquid interfaces of this circuit. Electrochemical reduction will take place at the grounded syringe needle (cathode), and oxidation reactions will (50) Atkins, P. Physical Chemistry, 6th ed.; W. H. Freeman & Co.: New York, 1998. (51) Kertesz, V.; Van Berkel, G. J. J. Mass. Spectrom. 2001, 36, 204-210.


4839

Figure 2. Acid-induced unfolding of cyt c in water/propanol (60/40 v/v) monitored by fluorescence spectroscopy (λex ) 280 nm, λem g 325 nm).

occur at the ESI needle (anode). With an external loop circuit present, charge-balancing reactions in the ESI capillary have to supply the electrons required for both currents, iesi and iext. Establishing an external loop circuit with iext . iesi is therefore expected to increase the rate of charge-balancing oxidation reactions in the ion source. In cases where the electrolytic oxidation of water (eq 1) is the main charge-balancing reaction, this effect should enhance the electrochemical acidification of the solution significantly. In the remainder of this study, we will explore this effect and its implications for the ESI mass spectra of cyt c. Acid-Induced Unfolding of cyt c. The heme protein cyt c is one of the most common model systems for studies on protein folding and unfolding.52-56 In aqueous solution, acid-induced unfolding of this protein occurs around pH 2.6.34,52,53 An electrochemical acidification during ESI will manifest itself most clearly in the transition region, where both the folded and the unfolded proteins are populated (analogous to the behavior of an acidbase indicator around its pKa). Furthermore, the ESI charge-state distribution will be most sensitive to changes in the proton concentration when unfolding occurs at near-neutral pH. Therefore, the unfolding transition of cyt c at pH 2.6 does not represent a good model system for studying the effects of electrochemical acidification. The addition of organic cosolvents destabilizes the native structure of cyt c and thus shifts the unfolding transition to higher pH.32,57 Figure 2 shows an unfolding curve of cyt c in a mixture of water and propanol (60/40 v/v), monitored by fluorescence spectroscopy. As a result of intramolecular quenching, the single tryptophan (Trp-59) in folded cyt c shows a low fluorescence intensity, whereas the unfolded protein is highly fluorescent.32,57,58 Figure 2 reveals that in the presence of 40% propanol the unfolding midpoint is around pH 5.6; at this pH, the solution represents a mixture of roughly equal amounts of folded and unfolded proteins. The acid concentration required to unfold the protein in the presence of 40% propanol is 3 orders of magnitude smaller than (52) Theorell, H.; Åkesson, A. J. Am. Chem. Soc. 1941, 63, 1812-1820. (53) Moore, G. R.; Pettigrew, G. W. Cytochromes c: Evolutionary, Structural and Physicochemical Aspects; Springer-Verlag: Heidelberg, 1990. (54) Roder, H.; Elo ¨ve, G. A.; Englander, S. W. Nature 1988, 335, 700-704. (55) Milne, J. S.; Xu, Y.; Mayne, L. C.; Englander, S. W. J. Mol. Biol. 1999, 290, 811-822. (56) Xu, Y.; Mayne, L.; Englander, S. W. Nat. Struct. Biol. 1998, 5, 774-778. (57) Bychkova, V. E.; Dujsekina, A. E.; Klenin, S. I.; Tiktopulo, E. I.; Uversky, V. N.; Ptitsyn, O. B. Biochemistry 1996, 35, 6058-6063. (58) Tsong, T. Y. Biochemistry 1976, 15, 5467-5473.


Figure 3. ESI mass spectra of cyt c at pH 5.6 in water/propanol (60/40 v/v) at a flow rate of 0.25 µL/min recorded with a platinum ESI capillary. The solution contained 1 mM KNO3: (A) Syringe needle not grounded; (B) syringe needle grounded. Notation: 12+ represents the gas-phase ion cyt c 12+, etc.

in the absence of organic cosolvents. The ratio of folded to unfolded protein in solution should therefore be highly sensitive to electrochemically induced pH changes. In the absence of any effects that alter the acidity of the solvent, the ESI mass spectrum recorded at pH 5.6 should display a bimodal charge-state distribution. Instrumental parameters such as differences in the sensitivity for ions of different m/z will influence the appearance of the mass spectrum. Regardless of these factors, electrochemical acidification will cause more of the proteins in solution to unfold, which should increase the relative intensity of highly charged protein ions in the spectrum. Effects of Water Oxidation the Mass Spectrum of cyt c, Obtained by Using a Platinum ESI Capillary. Figure 3A shows a mass spectrum of cyt c in an aqueous solution containing 40% (v/v) propanol at pH 5.6, electrosprayed from a platinum capillary at a flow rate of 0.25 µL/min. An electrolyte (1 mM KNO3) was added to increase the conductivity of the solution. The syringe needle was not connected to ground. The use of a platinum capillary implies that the electrolytic oxidation of water (eq 1) should be one of the major charge-balancing reactions.25 As expected on the basis of the fluorescence data discussed above, the spectrum displays a bimodal charge-state distribution. It has two maximums of similar intensity at 12+ and 8+ (n+ denotes gas-phase ions cyt cn+). On the basis of the experimentally measured current (itot ) iesi ≈ 60 nA), eq 3 predicts ∆pHmax ) -1.8 under these conditions. It is noted again that this calculation is based on a number of simplifying assumptions and that the actual pH change will often be less than predicted by eq 3. Figure 2 indicates that a pH change of 1.8 units should bring the protein out of the transition region. However, the observed bimodal charge-state distribution clearly shows that this is not the case. Interestingly, a similar ESI charge-state distribution was observed when the protein was sprayed in the absence of KNO3 under otherwise identical conditions, resulting in a current of ∼5 nA (which corresponds to ∆pHmax ) -0.8, data not shown). Therefore, it appears that the actual pH change in the presence of KNO3

Figure 5. ESI mass spectrum of cyt c at pH 5.6 in water/propanol (60/40 v/v) recorded with a platinum ESI capillary and a grounded syringe needle at a flow rate 50 µL/min.The solution contained 1 mM KNO3.

Figure 4. Intensities of two selected protein ions monitored as a function of time (upper trace, 13+; lower trace, 8+). Arrows indicate the time intervals during which the syringe needle was grounded. All the other conditions are as described for Figure 3.

is significantly smaller than predicted by eq 3. Possible reasons include the buffering capacity of the protein and of the acetic acid in the solution. Also the presence of easily oxidizable contaminants is likely to play a role.7 It will become clear later that the spectrum in Figure 3A does in fact indicate a certain degree of electrochemical acidification (see the discussion of Figures 6B and 8A). When the syringe needle was connected to ground under otherwise identical conditions, the measured current increased to ∼300 nA, which corresponds to ∆pHmax ) -2.5. The resulting spectrum (Figure 3B) shows a unimodal charge-state distribution with a maximum at 15+, indicating that virtually all of the proteins in solution are unfolded.28,31,32,34 The effects of grounding the syringe needle are also demonstrated in Figure 4, where the intensities of two selected ions, 13+ and 8+, are plotted as a function of time. These ions correspond to unfolded and folded protein molecules in solution, respectively. Grounding of the syringe needle (indicated by the arrows in Figure 4) is followed by a rapid increase of the 13+ ion intensity and causes the 8+ ion intensity to decrease. This effect is reversed when the grounding cable is disconnected. The data shown in Figures 3 and 4 confirm that the electrochemical acidification depends on the magnitude of itot. Increasing itot by closing the external loop circuit leads to a more pronounced acidification of the solution within the ESI capillary, which causes more of the proteins to unfold. According to eq 2, the electrochemical acidification of the protein solution should also depend on the solution flow rate vf. Figure 5 shows an ESI mass spectrum that was obtained at a flow rate of 50 µL/min. The resulting current (with the syringe needle grounded) was itot ≈ 870 µA. In other words, the solution flow rate under these conditions was 200 times higher than for the spectrum in Figure 3B, whereas the electron current drawn from the ESI needle only increased by a factor of ∼3. Under these conditions, eq 3 predicts ∆pHmax ) - 0.7. The resulting spectrum (Figure 5) shows a bimodal charge-state distribution. When the flow rate was reduced to 0.25 µL/min, the charge-state distribution changed back to that of Figure 3B. These data confirm that the

Figure 6. ESI mass spectra of cyt c at pH 5.6 in water/propanol (60/40 v/v) recorded with a platinum ESI capillary and a grounded syringe needle at a flow rate 0.25 µL/min. The solution contained 1 mM KI: (A) fused-silica capillary positioned flush with the ESI capillary (as indicated in Figure 1); (B) fused-silica capillary pulled 5 mm back.

solution flow rate is an important parameter that controls the extent of electrochemical acidification. It is noteworthy that the protein peaks obtained at an increased flow rate are significantly broadened due to the formation of K+ adducts. No such adducts were observed under the conditions of Figure 3B. This is consistent with previous studies that reported a higher “salt tolerance” of the ESI process at low flow rates.10, 59 Previous work has shown that it is possible to limit the extent of water oxidation by using an electrolyte that acts as a redox buffer, i.e., a substance that is more easily oxidized than water.25,60 Iodide is known to be an effective redox buffer for this purpose (overall reaction: 2I- f I2 + 2e-, E0 ) -0.53 V). Figure 6A shows an ESI mass spectrum of cyt c that was electrosprayed under conditions identical to those used in Figure 3B (i.e., grounded syringe needle, vf ) 0.25 µL/min), except that the solution contained 1 mM KI instead of 1 mM KNO3. The measured itot was ∼270 nA. Comparison of this spectrum with Figure 3B clearly shows that the electrochemical acidification of the solution is greatly suppressed, although the electron current drawn from the ESI needle is very similar. (59) Wilm, M.; Mann, M. Anal. Chem. 1996, 68, 1-8. (60) Moini, M.; Cao, P.; Bard, A. J. Anal. Chem. 1999, 71, 1658-1661.


4841

Figure 7. ESI mass spectrum of cyt c in water/propanol (60/40 v/v) recorded with a stainless steel ESI capillary at a flow rate 50 µL/min. The solution contained 1 mM KI and the syringe needle was not grounded; 0.75 mM HCl was added to the solution.

In its standard position, the fused-silica capillary was flush with the ESI capillary (see Figure 1). Generally, this positioning resulted in the most stable ESI MS signals and the lowest level of chemical noise. However, under these conditions, the contact between the analyte solution and the metal of the ESI capillary is limited to a small area at the tip of the sprayer. Electrochemical oxidation of a solute can only occur at this metal/liquid interface. Therefore, it appears that the extent of iodide oxidation in Figure 6A could be limited by the rate of iodide diffusion to the metal surface.7,16,61 To test this hypothesis, the area of the metal/liquid interface was increased by pulling the fused-silica capillary back by 5 mm. This drastically reduced the relative contribution of highly charged protein ions in the mass spectrum (Figure 6B). The measured itot remained constant around 270 nA. Pulling the fused-silica capillary back more than 5 mm did not decrease the contribution of highly charged protein ions any further. These observations strongly indicate that, under the conditions used for Figure 6A, the extent of iodide oxidation is indeed limited by mass transport, i.e., by the molecular diffusion of iodide ions to the metal surface.61 In Figure 6B, the oxidation of iodide appears to be the dominant charge-balancing reaction; i.e., this spectrum is essentially free of “electrochemical artifacts”. A simple calculation shows that ∼70% of the total I- ions in solution would have to be consumed to supply the measured itot entirely by iodide oxidation. In an additional experiment, an attempt was made to simulate the effects of water oxidation under conditions that strongly suppress electrochemical acidification. Hydrochloric acid at a concentration of 0.75 mM was added to the bulk solution. This corresponds to the proton concentration that is expected to accumulate under the conditions of Figure 3B (itot ) 300 nA, vf ) 0.25 µL/min), assuming that water oxidation is the only chargebalancing reaction. To minimize the effects of water oxidation, the flow rate vf was increased to 50 µL/min and 1 mM KI was added as a redox buffer. Furthermore, the solution was sprayed from a stainless steel capillary (the significance of this last point is addressed in the next section). An ESI mass spectrum of cyt c recorded under these conditions is shown in Figure 7. It shows a unimodal charge-state distribution with a maximum at 15+ that closely resembles the spectrum shown in Figure 3B. Effects of Water Oxidation on the Mass Spectrum of cyt c, Obtained by Using a Stainless Steel ESI Capillary. As pointed out earlier, iron oxidation can be the dominant charge(61) Van Berkel, G. J. J. Am. Soc. Mass Spectrom. 2000, 11, 951-960.


Figure 8. ESI mass spectra of cyt c at pH 5.6 in water/propanol (60/40 v/v) recorded with a stainless steel ESI capillary at a flow rate 0.25 µL/min. The syringe needle was grounded and the solution contained 1 mM KNO3. The data were measured (A) 1, (B) 2, and (C) 15 min after polishing the tip and the inside of the ESI capillary with a thin wire. The intensity of the spectra has been corrected for the different number of scans used.

balancing reaction in a stainless steel ESI capillary, at least in organic solvents such as methanol or acetonitrile and in the absence of easily oxidizable solutes.6,17 The metal of a stainless steel capillary can also act as a redox buffer in aqueous solution and therefore suppresses water oxidation to a certain extent.25 However, it is well known that water oxidation will still occur on the surface of a stainless steel electrode when the applied potential is high enough. A certain degree of metal dissolution takes place under these conditions, but the rate of this process is limited by a passivation layer on the electrode surface that consists of complex mixed-metal oxides.62 This is in line with the findings of a recent study60 that reported water oxidation at the sprayer of a capillary electrophoresis-ESI MS system. No differences were observed between electrodes made of platinum and stainless steel. It thus appears that the capacity of stainless steel capillaries to act as redox buffers during ESI has to be evaluated on a caseby-case basis. Figure 8 describes the results of an experiment where the interior tip region and the front face of a stainless steel ESI sprayer were carefully polished by rotating a thin wire inside the capillary. Subsequently, mass spectra were recorded under the conditions of Figure 3B (i.e., flow rate 0.25 µL/min, pH 5.6, 40% propanol, 1 mM KNO3, syringe needle grounded; itot ≈ 280 nA). Figure 8A shows a spectrum obtained after electrospraying from the polished ESI capillary for roughly 1 min. The high charge states in this (62) Sedriks, A. J. Corrosion of Stainless Steels, 2nd ed.; John Wiley & Sons: New York, 1996.

spectrum have low relative intensity, marking the presence of relatively little unfolded protein. This spectrum indicates only very little, if any, electrochemical acidification of the solution. After ∼2 min the intensity of highly charged protein ions has strongly increased, indicating a higher concentration of unfolded protein in the ESI capillary (Figure 8B). This trend continues until after ∼15 min no further changes are observed (Figure 8C). The data in Figure 8 clearly indicate a rapid decrease of the pH within the ESI capillary. This causes more of the protein to unfold and therefore changes the charge-state distribution of the observed spectra. It appears that polishing the capillary tip removes the passivation layer on the metal surface. Initially, this allows metal dissolution to occur. Under these conditions, the stainless steel capillary acts as a redox buffer as most of the charge-balancing electrons will be supplied by iron oxidation. Note the similarity of Figure 8A with the spectrum depicted in Figure 6B, where KI was added as redox buffer. As the passivation layer on the metal surface is reestablished, iron oxidation becomes more difficult; therefore, more and more electrons are supplied by water oxidation.62 It was not attempted to confirm the formation of a passivation layer on the metal surface by other methods. However, given the appearance of the spectra in Figure 8 there can be little doubt that this process plays a major role for the observed effects. A comparison of Figure 8C (stainless steel with passivation layer) and Figure 3B (platinum) shows that the spectra observed under these conditions are similar. The spectrum in Figure 3B is shifted to slightly higher charge states, which indicates a more acidic pH within the platinum ESI capillary. Iron oxidation might therefore still contribute as a charge-balancing reaction under the conditions of Figure 8C. This is in line with the observation of some corrosion at the tip of the stainless steel sprayer after ∼20 h of continuous operation. No signs of corrosion were observed for the platinum capillary. In an attempt to increase the redoxbuffering capacity of the stainless steel sprayer, the fused-silica capillary was pulled back several millimeters, similar to the experiment described in Figure 6B. However, this did not result in any significant changes of the observed ESI charge-state distribution (data not shown). CONCLUSIONS The experiments described in the this work provide clear evidence that the ESI mass spectra of proteins can be dramatically affected by electrolytic water oxidation. Electrochemical acidification can cause proteins in the ESI capillary to unfold, which shifts the observed ESI charge-state distribution to higher protonation states. Electrolytic acidification has to be taken into account for studies that employ the ESI charge-state distribution as a probe for conformational changes of proteins, especially in cases where unfolding occurs close to neutral pH. Also for studies on noncovalent ligand-protein complexes this is an important issue, as the stability of these complexes depends critically on the protein conformation and the solution pH.36 Electrochemical acidification can be suppressed by using protein solutions of low ionic strength and by not grounding the syringe needle. Both of these measures reduce the magnitude of itot and thus suppress water oxidation. The addition of redox buffers can also be beneficial. Acidification will be most pronounced at low flow rates, a factor that has to be considered when ESI sources that operate in the nanoliter per

minute regime are used.59 Only very minor pH changes are expected in cases where (bio)chemical processes are studied by time-resolved38,43 or stopped-flow63,64 ESI MS, as these techniques require flow rates on the order of 30-300 µL/min. The use of a pH buffer in the analyte solution will tend to suppress the effects of electrochemical acidification. However, addition of a buffer will also increase the conductivity of the solution and therefore the magnitude of itot. Therefore, it seems possible that pH buffers of low capacity will actually increase the extent of acidification, rather than decreasing it. In addition, the use of pH buffers in ESI MS is not straightforward, as solvent additives of this kind often lead to signal suppression and pronounced peak tailing.65 Electrochemical acidification is interesting from an analytical point of view, as it can increase the signal intensity significantly. Of all the ESI data shown in this study, the spectrum shown in Figure 3B is most strongly affected by electrochemical acidification; it shows the highest signal intensity and it has the highest signal-to-noise ratio. Acids are often added to the analyte solution in order to enhance the signal intensity in positive-ion ESI MS. However, for analytes that are not stable under acidic conditions, this strategy will not always be practical. In these cases, it might be possible to exploit electrochemical acidification to limit the acid exposure of the analyte to a very short time immediately before the spraying process takes place. The elements involved in the operation of a “standard” ESI source form a series circuit that represents a controlled-current electrolytic cell.6,7,16 An “external loop” is added to this circuit when the syringe needle is grounded. The resulting system consists of two electrolytic cells that share a common anode (i.e., the ESI capillary). The electrons produced by oxidation in the ESI capillary are consumed by two processes: (i) neutralization of positive charges emitted from the ion source (at the mass spectrometer, cathode 1) and (ii) electrochemical reactions at the syringe needle (cathode 2). Most likely the reduction of water, resulting in the generation of H2 and OH-, is the major electron-consuming reaction at cathode 2. Hydroxide anions produced in this reaction will flow down the fused- silica capillary and neutralize some of the protons generated during water oxidation. However, the data shown in this work indicate that this neutralization affects only a small fraction of the protons generated within the ESI capillary. Presumably the concentration of OH- formed at cathode 2 is relatively low, since the active electrode area inside the syringe needle is much larger than the active area of the anode (i.e., the tip of the ESI capillary). Adding an external loop to the ESI circuit can dramatically increase the electron current drawn from the ESI needle. It should be possible to use this effect for enhancing the formation of radical cations in cases where ESI is used for electrochemical ionization.7,13 Studies to test the viability of this concept are currently underway in our laboratory. ACKNOWLEDGMENT The authors thank David W. Shoesmith and Jamie Noe¨l for many helpful discussions regarding the electrochemical aspects (63) Kolakowski, B. M.; Simmons, D. A.; Konermann, L. Rapid. Commun. Mass Spectrom. 2000, 14, 772-776. (64) Kolakowski, B. M.; Konermann, L. Anal. Biochem. 2001, 292, 107-114. (65) Xu, N.; Lin, Y.; Hofstadler, S. A.; Matson, D.; Call, C. J.; Smith, R. D. Anal. Chem. 1998, 70, 3553-3556.


4843

of this work. We gratefully acknowledge the help and support of John Vanstone in the Departmental electronics shop. We also thank Robert K. Boyd and the NRC Institute of Marine Biosiences in Halifax, NS, for donating the Toby mass spectrometer to our laboratory. Financial support was provided by The Natural Sciences and Engineering Research Council of Canada (NSERC)

4844


and The University of Western Ontario. E.A.S. was recipient of a CAPES fellowship. Received for review May 14, 2001. Accepted July 30, 2001. AC010545R

Electrochemically Induced pH Changes Resulting in Protein Unfolding

Recommend Documents