Anal. Chem. 1999, 71, 5569-5573
Effective and Reproducible Capillary Electrophoretic Separation of Thiols under Conditions Where Exceptionally High Current Is Generated Eugenia V. Trushina,† Robert P. Oda,‡ Cynthia T. McMurray,† and James P. Landers*,§,|
Department of Pharmacology, Department of Biochemistry and Molecular Biology, Mayo Clinic and Foundation, Rochester, Minnesota 55905, and Department of Chemistry, Chevron Science Center, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, Department of Chemistry, McCormick Road, University of Virginia, Charlottesville, Virginia 22901
There is an escalating interest in the role of endogenous nitric oxide (NO) in biological systems and how this chemical regulates physiology in normal and disease states. In biological systems, the cellular concentration can be estimated, in the simplest form, by accounting for NO and its common metabolites, nitrate and nitrite. However, since NO is also known to interact with other chemical entities, such as thiols, it would be valuable to have a rapid qualitative assay that could account for thiol binding and S-N bond cleavage in the presence of different reducing agents. A separation buffer consisting of 10 mM phosphate, 10 mM HCl, and 250 mM KCl is shown to be adequate for the separation of glutathione, nitrosylated glutathione, and glutathione disulfide solubilized in 2 M HCl. The current observed under these separation conditions (249 µA at 11 kV) is extremely high by capillary electrophoresis (CE) standards, with a total power (current × voltage/capillary length) calculated to be in excess of 7 W/m. While this exceeds the ∼1.0 W/m recommended by previous studies as a maximum for CEbased separations, we demonstrate that effective CE separation of thiols can, in fact, be accomplished under these conditions with acceptable reproducibility, provided that buffer depletion issues are addressed. One of the attributes of capillary electrophoresis (CE) that makes it an attractive analytical technique is its versatility.1,2 Electrophoretic separation of analytes with diverse chemical character is possible in a fused silica capillary where ionic species as small as inorganic ions and as large as proteins can be separated.1 Prior to the advent of CE, the smallest molecules to * Corresponding author: (mail) Department of Chemistry, McCormick Road, University of Virginia, Charlottesville, VA 22901; (phone) 804-243-8658; (fax) 804243-8852; (e-mail)
[email protected]. † Department of Pharmacology, Mayo Clinic and Foundation. ‡ Department of Biochemistry and Molecular Biology, Mayo Clinic and Foundation. § University of Pittsburgh. | University of Virginia. (1) Jorgenson, J.; Lukacs, K. Anal. Chem. 1981, 53, 1298. (2) Landers, J. P. Clin. Chem. 1995, 41, 495-509. 10.1021/ac990505b CCC: $18.00 Published on Web 11/09/1999
© 1999 American Chemical Society
be separated in an electrophoresis-based format were amino acids via high-voltage paper electrophoresis.3 Capillary-ion electrophoresis (CIE) has taken CE of small molecules far beyond that of amino acids, having been shown to be effective for a variety of small ions.4,5 Currently, there is an intense effort to understand the chemistry of nitric oxide (NO) in biological systems and how this endogenous chemical regulates physiology in normal and pathological states.6 To do so, analytical methods must be devised to accurately measure the most important nitric oxide (NO) metabolites produced in vivo. In a previous report,7 we demonstrated that capillary-ion electrophoresis (CIE) could be exploited for understanding the reduction of nitrate and nitrite to NO in the presence of reducing agents. In addition to CIE being used for rapid quantification of nitrate and nitrite anions in biological matrixes, the optimal conditions for reduction of nitrite and nitrate to nitric oxide were determined, which can potentially be applied to more sensitive detection methods (e.g., chemiluminescence). NO is not only known to be oxidized to nitrate and nitrite, but also has a proclivity for binding to thiols and dithiols.8 While reports in the literature have detailed conditions for the capillary electrophoretic analysis of thiols,9,10 they do not directly address the ability to analyze for these compounds in sample matrixes such as 2 M HCl. Since we had devised conditions for studying the reduction of nitrite and nitrate to nitric oxide as a means of measuring the most important NO metabolites produced in vivo,7 we sought to extend the use of this analytical technique to qualitatively evaluate the binding of NO to thiols as well as S-N bond cleavage. In this report, we demonstrate that the electrophoretic resolution of glutathione (GSH), nitrosylated glutathione (3) Wieland, T.; Fischer, E. Naturwissenschaften 1948, 35, 29. (4) Jones, B. Handbook of Capillary Electrophoresis, 2nd ed., Landers, J. P., Ed.; CRC Press: Boca Raton, FL, 1996; pp. (5) Jones, W. R. J. Chromatogr. 1993, 640, 387. (6) Moncada, S.; Palmer, R. M. J.; Higgs, E. A. Pharmacol. Rev. 1991, 43, 109-142. (7) Trushina, E. V.; Oda, R. P.; Landers, J. P.; McMurray, C. T. Electrophoresis 1997, 18, 1890. (8) Stamler, J. S.; Singel, D. J.; Loscalso, J. Science (Washington, D.C.) 1992, 258, 1893-1902. (9) Russell, J.; Rabenstein, D. L. Anal. Biochem. 1996, 242, 136-144. (10) Stamler, J. S.; Loscalzo, J. Anal. Chem. 1992, 64, 779-785.
Analytical Chemistry, Vol. 71, No. 24, December 15, 1999 5569
(GSNO), and glutathione disulfide (GSSG) in a 2 M HCl sample matrix is feasible provided the separation buffer contains 250 mM KCl. Despite the very high system current (>220 µA) observed with moderate applied voltages (10 kV), translating to a power of ∼6 W/m of capillary, reproducible separations can be obtained using a buffer-replenishment approach. This provides a qualitative assay for nitrosylated thiols that may be applied to biological matrixes, but demonstrates that conditions considered to be outside the acceptable limits for effective CE can, in fact, yield acceptable and useful separations. EXPERIMENTAL SECTION Materials. Phosphoric acid (H3PO4) was purchased from GFS Chemicals (Powell, OH). Glutathione and glutathione disulfide were purchased from Acros (Pittsburgh, PA). Sodium nitrite (NaNO2) and O-phosphorylethanolamine (o-PEA) were from Sigma (St. Louis, MO). Potassium chloride was from Amresco (Solon, OH). All other chemicals were of analytical-grade or better. Double-distilled water was used for preparation of all aqueous solutions of the reagents. Sample Preparation. All solutions were prepared fresh prior to the CE separation. GSNO was prepared by reacting equimolar (0.1 M) concentrations of GSH with NaNO2 in 1 M HCl.11 The concentration of nitrosothiol was assessed by Saville reaction12 and UV-vis spectrophotometry using absorbance at 335 nm and the reported extinction coefficient.13 GSH and GSSG were dissolved in 10 mM HCl to obtain a final concentration of 0.1 M. Before CE separation, GSNO, GSH, and GSSG were mixed and then dissolved to the desired concentration (5-10 mM) in water (final pH 2.5) or 2 M HCl (final pH 0). The mixture was placed in a sample vial thermostated at 37 °C and immediately analyzed by CE. Capillary Electrophoresis. CE separation was carried out on a Beckman P/ACE System 2100 equipped with a monochromatic UV detector and sample temperature control tray (to maintain samples at 37 °C). An IBM 55SX computer utilizing System Gold software (version 8.1) was used for instrument control and data collection. All peak information (migration time and integrated peak areas) was obtained through the System Gold software. The separation was performed in bare-silica capillaries (Polymicro Technologies, Phoenix, AZ) having a 50-µm i.d. and a 37-cm length (30 cm to the detector). New capillaries were conditioned with a 20-column volume rinse with each of 1.0 M NaOH, water, and finally the separation buffer. Four main CE separation buffers were evaluated in this study: buffer 1, 10 mM phosphate; buffer 2, 10 mM phosphate with 10 mM HCl; buffer 3, 10 mM phosphate, 10 mM HCl, and 0.75 M o-PEA; and buffer 4, 10 mM phosphate, 10 mM HCl, and 0.25 M KCl. All buffers were adjusted to pH 2.5 with 0.1 M NaOH. For a typical analysis, the following method was used: a 3-column volume rinse (20 psi) with running buffer, a low-pressure injection (0.5 psi) of sample, 1-s low-pressure (0.5 psi) injection of separation buffer, separation at 9-15 kV (constant voltage), a 3-column volume wash (20 psi) with 0.1 M H3PO4 followed by a 3-column volume rinse (20 psi) with running buffer. (11) Loscalzo, J. J. Clin. Invest. 1985, 76, 703-708. (12) Saville, B. Analyst (Cambridge, U.K.) 1958, 83, 670-672. (13) Stamler, J. S.; Feelisch, M. Methods in Nitric Oxide Research; Feelisch, M., Stamler, J. S., Eds.; John Wiley & Sons, Ltd.: New York, 1996; pp 521538.
5570 Analytical Chemistry, Vol. 71, No. 24, December 15, 1999
Figure 1. The effect of sample matrix pH on separation of glutathione disulfide (GSSG), glutathione (GSH), and nitrosylated glutathione (GSNO). (A) A mixture containing 5 mM GSSG, 5 mM GSH, and 10 mM GSNO in HCl (pH 2.5) separated in 10 mM phosphate buffer, pH 2.5; detection at 200 nm. Separation conditions: 50 µm × 37 cm bare silica capillary. Applied voltage 15 kV, current - 22 µA. Sample injection 5 s at 0.5 psi (3447 Pa). Capillary thermostated at 20 °C. Sample vial thermostated at 37 °C. (B) A mixture containing 5 mM GSSG, 10 mM GSH, and 10 mM GSNO in 2 M HCl (pH 0) separated in 10 mM phosphate, 10 mM HCl; pH 2.5. Applied voltage 15 kV, current - 22 µA. Other separation conditions the same as in (A). Peaks: 1, GSSG; 2, GSH.
The samples were injected for 5 s. Polarity was such that the inlet was the anode and the outlet was the cathode. The capillary temperature was thermostated at 20 °C, and detection was by absorbance at 200 nm. RESULTS AND DISCUSSION While the conditions reported by Russell and Rabenstein9 demonstrate that thiols/dithiols in a neutral pH sample matrix could be resolved using 100 mM phosphate buffer at pH 2.3, we found that these conditions led to excessively long analysis times (>60 min using a 37-cm capillary) and, consequently, were not adequate for our purposes. When GSSG, GSH, and GSNO were dissolved in an acidic sample matrix (pH 2.5) as described by Stamler et al.10 and electrophoresed in a 20 cm × 25 µm capillary containing 10 mM phosphate/0.01N HCl (pH 2.5), the separation of GSSG from the other components was observed within 20 min, but GSH and GSNO were unresolved (data not shown). When the effective capillary length was extended to 37 cm and the separation buffer simplified, all three analytes were adequately resolved, albeit with a slightly longer analysis time (Figure 1A). Our previous work7 differed from these studies in that the developed CIE method was capable of analysis of samples of nitrite and nitrate in a 2 M HCl sample matrix having a pH of 0.0. This was accomplished using a detergent-based separation buffer containing 200 mM lithium chloride, which minimized the effect of high-ionic-strength sample matrixes. Since the goal of the present study was to evaluate other forms of bound NO that existed (e.g., nitrosothiols) under the same reducing conditions, it was essential that the method be capable of resolving analytes in a 2 M HCl (pH 0, 37 °C) sample matrix. Figure 1B shows the deleterious effects on the CE separation when GSSG, GSH, and GSNO were dissolved in a 2 M HCl sample matrix and electrophoresed with the separation buffer described by Stamler et al.10 (10 mM phosphate/10 mM HCl, pH 2.5). Poor peak shape with severe fronting was clear with GSSG, and GSNO was not observed even after electrophoresis for 52 min.
Figure 2. The separation of glutathione disulfide (GSSG), glutathione (GSH), and nitrosylated glutathione (GSNO) when the buffer is modified to accommodate a 2 M HCl sample matrix. (A) A mixture containing 5 mM GSSG, 10 mM GSH, and 10 mM GSNO in 2 M HCl separated in 10 mM phosphate, 10 mM HCl, 750 mM o-PEA; pH 2.5. Other conditions: 50 µm × 37 cm bare silica capillary; detection at 200 nm. Applied voltage 10 kV, current -17 µA. (B) Sample injection 5 s at 0.5 psi (3447 Pa). Capillary thermostated at 20 °C. Sample vial thermostated at 37 °C. (B) A mixture containing 5 mM GSSG, 10 mM GSH, and 10 mM GSNO in 2 M HCl separated in 10 mM phosphate, 10 mM HCl, 250 mM KCl; pH 2.5. Applied voltage 10 kV, current -220 µA. Peaks: 1, GSSG; 2, GSH; 3, GSNO.
Increasing the separation buffer-sample matrix compatibility was initially attempted by augmenting the ionic strength of the 10 mM phosphate/10 mM HCl separation buffer with 750 mM o-phosphorylethanolamine (o-PEA). o-PEA was effective for this purpose since its zwitterionic character allows for an increase in ionic strength without a proportional increase in current. Figure 2A shows an exemplary separation of GSSG, GSH and GSNO (2 M HCl sample matrix) under these conditions. While all three analytes could be observed, this separation was associated with poor peak shape (particularly for GSSG and GSH), the appearance of additional unidentified system peaks and an erratic baseline. The erratic baseline may be due to fluctuations in the current reflecting thermostatic cycling at the limits of temperature control. To better match the high chloride concentrations of the sample matrix, the 750 mM o-PEA was substituted with 250 mM KCl. As can be seen in Figure 2B, all three analytes could be resolved and detected under these conditions with a reasonably stable baseline and good peak shape. However, the improved compatibility between the separation buffer and the sample matrix was not obtained without a cost. At a moderate applied potential of 10 kV, the current associated with this separation plateaued at ∼220 µA (5.95 W/m), an exceptionally high current by CE standards that most likely results from the high chloride concentration. Electrolytes such as chloride, citrate, and sulfate are known to have a high mobility and, as such, generate high system current and can induce excessive Joule heating at elevated concentrations. Excessive Joule heating is avoided in CE since it can cause irreproducible sample injection,14,15 intracapillary outgassing or bubble generation,16 and band broadening,17as well as affect sample stability15,16 and quantification.14,16,18 The power applied to a capillary electrophoretic system is defined by the Ohm’s Law relationship as follows: (14) McCormick, R. M. Anal. Chem. 1988, 60, 2322. (15) Rush, R. S.; Cohen, A. S.; Karger, B. L. Anal. Chem. 1991, 63, 1346. (16) Nelson, R. J.; Paulus, A.; Cohen, A. S.; Guttman, A.; Karger, B. L. J. Chromatogr. 1989, 480, 111. (17) Gobie, W. A.; Ivory, C. F. J. Chromatogr. 1990, 516, 191. (18) Lookabaugh, M.; Biwas, M.; Krull, I. S. J. Chromatogr. 1991, 549, 357.
Power (mW/m) )
current (µA) × voltage (kV) capillary length (m)
Application of 25 kV to a buffer-filled capillary 1 m in length with a system current of 30 µA is associated with an applied power of 0.750 W or 0.863 W per meter of capillary (W/m). As indicated above, system current is dependent on the applied voltage, capillary dimensions, and the electrolytes in the capillary and their concentration. As the ionic strength increases, particularly with high-mobility electrolytes, such as chloride, citrate, and sulfate, the current is elevated, increasing the total power associated with the system. The power limits for optimal CE have been defined by Sepaniak and Cole19 as
