Anal. Chem. 2001, 73, 2898-2902
Speciation and Identification of Organoselenium Metabolites in Human Urine Using Inductively Coupled Plasma Mass Spectrometry and Tandem Mass Spectrometry Tiffany H. Cao, Rita A. Cooney, Michelle M. Woznichak, Sheldon W. May, and Richard F. Browner*
School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia 30332-0400
A method for speciation and identification of organoselenium metabolites found in human urine samples using high performance liquid chromatography/inductively coupled plasma mass spectrometry (HPLC/ICPMS) and tandem mass spectrometry (MS/MS) is described. Reversed-phase chromatographic separation was used for sample fractionation with the ICP-MS functioning as an element-selective detector, and six distinct seleniumcontaining species were detected in a human urine sample. Fractions were then collected and analyzed using a triple quadrupole mass spectrometer with electrospray ionization and collision-induced dissociation to obtain structural information. The first two fractions were identified specifically as selenomethionine and selenocystamine, estimated to be present at approximately 11 and 40 ppb, respectively. To the best of our knowledge, this is the first time these two metabolites have been positively identified in human urine. The epidemiology, biology, and biochemistry of selenium are subjects of intense current interest, particularly from the perspective of public health.1,2 The role of selenium as a dietary antioxidant has been recognized for some time,3-7 and more recently its role as an anticarcinogenic agent has become apparent.8-12 Selenium dietary deficiency has been linked to heart diseases,13 arthritis,14 * Phone: 404-894-4020. Fax: 404-894-7452. E-mail: rick.browner@ chemistry.gatech.edu. (1) May, S. W. Expert Opin. Invest. Drugs 199, 8, 1017-1030. (2) May, S. W.; Pollock, S. H. Drugs 1998, 56, 959-964. (3) Walter, R.; Schwartz, I. L.; Roy, J. Ann. N.Y. Acad. Sci. 1972, 192, 175180. (4) Leibovitz, B.; Hu, M.; Tappel, A. L. J. Nutr. 1990, 120, 97-104. (5) Caldwell, K. A.; Tappel, A. L. Biochemistry 1964, 3, 1643-1647. (6) Olcott, H. S.; Brown, W. D.; Van der Veen, J. Nature 1961, 191, 12011202. (7) Hamilton, J. W.; Tappel, A. L. J. Nutr. 1963, 79, 493-502. (8) Medina, D. J. Am. Coll. Toxicol. 1986, 5, 21-27. (9) Ip, C.; Hayes, C. J. Am. Coll. Toxicol. 1986, 5, 7-19. (10) Vernie, L. N. Biochim. Biophys. Acta 1984, 738, 203-217. (11) Schrauzer, G. N. Biol. Trace Elem. Res. 1992, 33, 51-62. (12) Clark, L. C.; Combs, G. F., Jr.; Turnbull, B. W.; Slate, E. H.; Chalker, D. K.; Chow, J.; Davis, L. S.; Glover, R. A.; Graham, G. F.; Gross, E. G.; Krongrad, A.; Lesher, J. L., Jr.; Park, H. K.; Sanders, B. B., Jr.; Smith, C. L.; Taylar, R. J. Am. Med. Assoc. 1996, 276, 1957-1963. (13) Huttnen, J. K. Biomed. Environ. Sci. 1997, 10, 220-226. (14) Kose, K.; Dogan, P.; Kardas, Y. Biol. Trace Elem. Res. 1996, 53, 51-56.
2898 Analytical Chemistry, Vol. 73, No. 13, July 1, 2001
cancer,15,16 and recently, AIDS.17 A recent large-scale human study has also confirmed that selenium supplementation reduces the incidence of prostate, colorectal, and lung cancer.12 Although large selenium-containing molecules, such as seleno-enzymes and selenium-binding proteins, have been a primary source of study for several years,18-21 it is now clear that the antitumorigenic effects of selenium must arise, at least in part, from mechanisms involving low-molecular-weight selenium metabolites;22,23 however, the pathways and mechanisms of selenium metabolism that lead to these beneficial effects are poorly understood. Therefore, the overall objective of our research has been to obtain structural information regarding the selenium-containing compounds formed in human metabolism. The speciation of selenium in biological samples has presented a major unresolved challenge to analytical chemistry for several years, primarily as a result of extremely low total concentrations of selenium species observed in human urine,24,25 and the inadequacy of conventional approaches for chromatographic separation and structure determination. Additionally, the lack of readily available authentic selenium standards and the known high thermal lability of many organoselenium compounds26-28 have complicated the process of selenium identification in complex biological matrixes. From the perspective of mass spectrometry, electrospray ionization is an effective “soft” ionization approach, but the high salt content of urine samples suppresses analyte ionization and reduces the already low signal intensity even further. Several recently published studies have demonstrated that effective separation of several low-molecular-weight selenium (15) Parnham, M. J.; Grap, E. Biochem. Pharmacol. 1987, 36, 3095-3102. (16) Fox, J. M. Methods Find. Exp. Clin. Pharmacol. 1992, 14, 275-287. (17) Baum, M. K.; Shor-Posner, G.; Lai, S. J. Acquired Immune Defic. Syndr. Hum. Retroviol. 1997, 15, 370-374. (18) Beck, M. A.; Shi, Q.; Morris, V. C. Nat. Med. 1995, 1, 433-436. (19) Stadtman, T. C. Annu. Rev. Biochem. 1996, 65, 83-100. (20) Ursini, F.; Bindoli, A. Chem. Phys. Lipids 1987, 44, 255-276. (21) Burk, R. F.; Hill, K. E. Bioessays 1999, 21, 231-237. (22) Ganter, H. E. Carcinogenesis 1999, 20, 1657. (23) Combs, G. F.; Gray, W. P. Pharmacol. Therap. 1998, 79, 179. (24) Cooney, R. Ph.D. Thesis, Georgia Institute of Technology, 1999. (25) Alaejos, M. S.; Romero, C. D. Clin. Chem. 1993, 39, 2040. (26) Shou, W. Z.; Woznichak, M. M.; May, S. W.; Browner, R. F. Anal. Chem. 2000, 72, 3266. (27) Welz, B.; Melcher, M.; Neve, J. Anal. Chim. Acta 1984, 165, 131. (28) Shou, W. Z. Ph.D. Thesis, Georgia Institute of Technology, 1999. 10.1021/ac0100244 CCC: $20.00
© 2001 American Chemical Society Published on Web 05/22/2001
species is possible.29-36 A limitation of these works was that the identification of the inorganic or organic selenium species was based solely on retention time comparisons with reference compounds and not on actual structural data obtained via mass spectrometry. However, retention time match does not unequivocally confirm chemical structure. Without mass spectral structure confirmation,37 only partial identification of the selenium species that are present in biological samples can me accomplished. Several workers38-43 have attempted to obtain the mass spectral data of selenium compounds in enriched and natural samples but did not report (except for Casiot et al.) MS/MS spectra of product ions for structural confirmation. To unequivocally confirm the structure of a given selenium-containing metabolite, it is essential to obtain positive identification of the organic part of the molecule. We now report the successful speciation of several organoselenium compounds found in human urine samples. Through the use of high performance liquid chromatography/inductively coupled plasma mass spectrometry (HPLC/ICP-MS) with an oscillating capillary nebulizer (OCN) interface,44 followed by structural determination using an electrospray tandem mass spectrometer, we have successfully separated and identified two selenium metabolites, selenomethionine and selenocystamine, in human urine. The detection limits for both selenium compounds in urine were estimated on the basis of a signal-to-noise ratio of 3:1, to be 10 µg/L. This is the first time, to our knowledge, that these two selenium metabolites have ever been detected and positively identified in human urine. EXPERIMENTAL SECTION Chemicals and Samples. All solvents were of HPLC grade. Acetic acid was ACS grade (Fisher Scientific; Atlanta, GA). SelenoDL-methionine and selenocystamine were purchased from Sigma Chemicals (St. Louis, MO). Standard solutions were prepared in 50:50 methanol:OPTIMA trace metal grade water (Fisher Scien(29) Gammelgaard, B.; Jessen, K.; Kristensen, F. H.; Jons, O. Anal. Chim. Acta 2000, 404, 47-54. (30) Zheng, J.; Ohata, M.; Furuta, N.; Kosmus, W. J. Chromatogr. A 2000, 874, 55-64. (31) Kotrebai, M.; Tyson, J. F.; Block, E.; Uden, P. C. J. Chromatogr. A 2000, 866, 51-63. (32) Sutton, K. L.; Ponce de Leon, C. A.; Ackley, K. L.; Sutton, R. M. C.; Stalcup, A. M.; Caruso, J. A. Analyst 2000, 125, 281-286. (33) Shiobara, Y.; Ogra, Y.; Suzuki, K. Analyst 1999, 124, 1237-1241. (34) Gonzalez-LaFuente, J. M.; Dlaska, M.; Gernandez-Sanchez, M. L.; SanzMedel, A. J. Anal. Atom. Spectrom. 1998, 13, 423-429. (35) Bird, S. M.; Ge, H.; Uden, P. C.; Tyson, J. F.; Block, E.; Denoyer, E. J. Chromatogr. A 1997, 789, 349-359. (36) Olivas, R. M.; Donard, O. F. X. J. Anal. Atom. Spectrom. 1996, 11, 11711176. (37) Li, L.; Campell, D.; Bennett, P.; Henion, J. Anal. Chem. 1996, 68, 33973404. (38) Kotrebai, M.; Birringer, M.; Tyson, J., Block, E.; Uden, P. Analyst 2000, 125, 71-78. (39) Kotrebai, M.; Birringer, M.; Tyson, J.; Block, E.; Uden, P. Anal. Commun. 1999, 6, 249-252. (40) Casiot, C.; Vacchina, V.; Chassaigne, H.; Szpunar, J.; Potin-Gautier, M.; Lobinski, R. Anal. Commun. 1999, 36, 77-80. (41) Kotrebai, M.; Bird, S. M.; Tyson, J. F.; Block, E.; Uden, P. C. Spectrochim. Acta Part B 1999, 54, 1573-1591. (42) Michalke, B.; Schramel, O.; Kettrup, A. Fresenius J. Anal. Chem. 1999, 363, 456-458. (43) Crews, H.; Clarke, P.; Lewis, J.; Owen, L.; Strutt, P.; Izquierdo, A. J. Anal. Atom. Spectrom. 1996, 11, 1177-1182. (44) U.S. Patent No. 5725153, issued March 10, 1998.
tific; Atlanta, GA). Urine samples were collected from an adult male for four consecutive days after a single intake of 400 µg of seleno-DL-methionine supplement (Nutralite, Solgar Vitamin and Herb Company Inc.; Leonia, NJ). Sample Preparation. Urine samples were stored at -70 °C in precleaned Teflon containers to minimize contamination or adsorption losses. Urine samples were preconcentrated because organoselenium compounds are known to be present in urine at a very low level (total concentration, 20-200 µg/L24,25). A known volume, between 20 and 100 mL, of the original urine sample was lyophilized using a freeze-dry system (Labconco). The lyophilized urine sample was then reconstituted to 1:4 with OPTIMA water. The reconstituted samples were filtered through a 0.2 µm syringe filter. The samples were then directly injected into the LC system without any further sample manipulation. HPLC Conditions. Separation was carried out on a housepacked Hypersil BDS C18 column (2 × 100 mm, 5-µm particle size). The mobile phase gradient was generated using (A) OPTIMA water with 2% acetic acid and (B) 90:10 methanol:2propanol. Injection volume was 50 µL, mobile phase flow rate was 0.1 mL/minute, and a gradient running from 1 to 55% B was used in all runs. HPLC/ICP Fractionation. Preconcentrated urine samples were first separated by HPLC and run through the ICP-MS (Perkin-Elmer SCIEX ELAN 5000) to obtain retention times of all selenium-containing compounds. All runs were made in triplicate. Then, sample fractions were collected using a split valve with 1% of the effluent going to the ICP-MS and 99% going to sample collection. With the ICP-MS as the element-selective detector, fraction collection began as soon as the ICP-MS selenium signals (74Se, 77Se, 78Se, 82Se) appeared. Mass Spectrometry. All inorganic mass spectra were obtained using a Perkin-Elmer SCIEX ELAN 5000 ICP-MS instrument, and organic mass spectra were obtained using a SCIEX API III biomolecular mass analyzer (Perkin-Elmer) used in MS/MS positive ion mode. The collision energies for the dissociation of selenomethionine and selenocystamine ions were 25 and 35 eV, respectively. RESULTS AND DISCUSSION All of the data presented in this paper were obtained using urine samples collected on day 1 of the experiment. The target selenium-containing species were initially unidentified but known to be present at quite low concentrations; therefore, all of the samples were preconcentrated by 4× (see sample preparation section) before LC separation. The preconcentration procedure was found to have no effect on the identities or relative concentrations of peaks in the speciation profile. The OCN used in the ICPMS part of this study was helpful in both optimizing analyte transport in the interface and minimizing organic solvent loading in the ICP.45 This is a direct consequence of several properties of the nebulizer. The device produces an extremely fine aerosol from the chromatographic effluent, with a mean drop size < 2 µm, substantially smaller than produced by conventional nebulizers.46 (45) Wang, L.; May, S. W.; Browner, R. F.; Pollock, S. H. J. Anal. Atom. Spectrosc. 1997, 11, 1137-1146. (46) Wang, L.; Tucker, A.; May, S. W.; Browner, R. F. Anal. Chem. 2001, submitted.
Analytical Chemistry, Vol. 73, No. 13, July 1, 2001
2899
Figure 1. Separation of selenium-containing compounds in concentrated (4×) Se-supplemented human urine using a house-packed Hypersil BDS C18 (2 × 100 mm, 5 µm) column with a flow rate of 0.1 mL/minute (see HPLC Conditions Section for experimental conditions). No significant signals of 35Cl and 79Br were observed.
As a consequence of the small mean drop size, solvent evaporated spontaneously and rapidly, without the need for application of an external heat source. Consequently, only dry, solvent free nanoparticles reached the plasma. Moreover, the device can be run efficiently at low effluent flow rates, down to < 1 µL/min, which allows for effluent splitting between the MS and sample fractionation system without undue signal loss. Figure 1 shows an ICP-MS chromatogram of a 50-µL urine sample (lyophilized and reconstituted) with four selenium isotopes, 74Se, 77Se, 78Se, 82Se, being monitored simultaneously. The most abundant isotope of selenium, 80Se, could not be used because of high isobaric interference from 40Ar2+ in the plasma. Nevertheless, the signals of all four of the isotopes, 74Se, 77Se, 78Se, 82Se, in each eluting band confirmed the presence of selenium in the urine samples. Possible isobaric effects from 40Ar37Cl and 1H81Br can also occur at masses 77 and 82. However, no signal above baseline was observed for the two 35Cl and 79Br ions, thus proving the absence of Cl and Br in the system. From the ICP-MS chromatogram, six selenium-containing compounds were observed in the examined urine sample, and the mass spectral identification of the first two fractions is presented in this paper. From Figure 1, the retention times of the first two fractions, 250 s and 350 s, matched those of authentic samples of selenoDL-methionine and selenocystamine (Figure 2). Therefore, it was suspected that these two fractions contained selenocystamine and selenomethionine, respectively. However, because ICP-MS is only an element-selective detector, retention time match does not guarantee a match of chemical structure. Consequently, the production of structurally significant mass spectra of the collected fractions was attempted, and selenocystamine and selenomethionine standards were used to provide reference mass spectra. In general, structural mass spectra of low-molecular-weight organoselenium compounds can be obtained using electron impact ionization (EI), chemical ionization (CI), or electrospray ionization MS/MS; however, excessive fragmentation of low-molecularweight organoselenium compounds resulted in a low response factor for EI and CI. Figure 3 shows an EI spectrum of selenomethionine obtained in positive ion mode using a VG Instruments 70-SE double-focusing magnetic sector mass spectrometer. A direct injection probe inlet and a 70 eV source were used. The ion source temperature was ramped from 30 °C to 400 °C, and the resulting mass spectrum was obtained under conditions giving maximum total ion current. The level of selenomethionine 2900 Analytical Chemistry, Vol. 73, No. 13, July 1, 2001
Figure 2. HPLC/ICP-MS chromatogram of (a) 500 ppm selenocystamine standard and (b) 150 ppm L-selenomethionine standard in 50:50 MeOH:H2O (same experimental conditions as in Figure 1).
Figure 3. Electron impact ionization mass spectrum of 250 µg/L selenomethionine standard. Standard solution was prepared in 80: 20 acetonitrile:water, allowed to evaporate, then put on the direct insertion probe.
required to obtain useful EI signals was in the range of 350-400 µg/L, which is too high to be useful for our present study. This level is about 10-fold higher than the total Se concentration found in a typical urine sample.24,25 For this reason, electrospray MS/ MS was preferred to provide soft ionization followed by collisioninduced dissociation (CID) to obtain structural information of the selenium-containing species detected in the examined urine samples. Shown in Figure 4 are the MS/MS spectra of selenomethionine and selenocystamine standards. The molecular peaks [M + H]+ of selenomethionine and selenocystamine at 198 and 249, respectively, clearly show the expected isotope patterns for a single Se atom in selenomethionine and for two Se atoms in selenocystamine. In contrast to the ICP-MS experiments, the 80Se isotope was used in all of the MS/MS experiments in order to maximize the signals in product ion scans, because clearly isobaric interference from 40Ar2+ was not an issue here. The precursor ion of selenocystamine (m/z 249) produced three prominent product ions at m/z 204, 124, and 44. The precursor ion of selenomethionine (m/z 198) also produced three prominent product ions at m/z 181, 152,
Figure 5. Schematic diagram of a MRM scan: Q1, 1st quadrupole; Q2, collision cell; Q3, 3rd quadrupole.
Figure 6. (a) MRM signal for selenocystamine in urine fraction 1 and (b) relative intensities of product ions for the selenocystamine standard and urine fraction 1. Figure 4. MS (a) and MS/MS (b) scans of 500 ppb selenocystamine standard; MS (c) and MS/MS (d) scans of 150 ppb selenomethionine standard. All standard solutions were prepared in 50:50 MeOH:H2O.
and 109 (see Figure 3 for product ion assignment47). The product ion at m/z 109 was unsuitable because of a high background signal observed in this m/z region originating from an unknown source. These precursor-product ion transitions were used in the multiple (47) Dookeran, N.; Yalcin, T.; Harrison, A. G. J. Mass Spectrosc. 1996, 31, 500508.
reaction monitoring (MRM) to monitor for selenomethionine and selenocystamine in the HPLC-collected urine fractions. The MRM mode, in which one or more transitions of precursor to product ions are monitored in tandem, is generally used for structural confirmation. A MRM scan gives a signal only if the specified precursor and product ions are detected concurrently by both quadrupoles [Q1 (precursor ion) and Q3 (product ion) in Figure 5]. The MRM mode was preferred over full-scan modes because it required less sample volume (20-50 µL vs 300-500 µL for fullAnalytical Chemistry, Vol. 73, No. 13, July 1, 2001
2901
Figure 7. (a) MRM signal for selenomethionine in urine fraction 2 and (b) relative intensities of product ions for the selenomethionine standard and urine fraction 2.
scan modes). In addition, MRM has been shown to typically increase the sensitivity by a factor of 50-100 when compared to full-scan modes of operation.48 In the MRM mode, the collected urine fractions (20 µL, fraction 1 or 2 in Figure 1) were introduced to the mass spectrometer through an injection loop by passing the LC column. The resulting (48) Vishwanathan, K.; Tackett, R. L.; Stewart, J. T.; Bartlett, M. G. J. Chromatogr. B 2000, 748, 157-166.
2902
Analytical Chemistry, Vol. 73, No. 13, July 1, 2001
MRM scans are shown in Figures 6a and 7a. For the first collected urine fraction, three transitions, m/z 249 to 44, 249 to 124, and 249 to 204, were monitored (Figure 6a), and all three of these transitions gave responses that peaked simultaneously, thus confirming the presence of selenocystamine. Similar results were observed for the second collected fraction, with the two transitions m/z 198 to 181 and 198 to 152 confirming the presence of selenomethionine (Figure 7a). To obtain the relative abundance of product ions in all transitions, the average of scans from the MRM signal in Figures 6a and 7a was acquired, and the results are shown in Figures 6b and 7b (solid line), which exhibit the precursor-to-product patterns characteristic of selenocystamine and selenomethionine. The difference in the relative intensity of each product ion compared to that obtained from the standards (broken line) was between 6% and 8% for both compounds. According to Henion37 et al., the acceptance criteria for positive MS/MS identification and confirmation are (i) HPLC retention time reproducibility (( 2%), (ii) at least two or preferably three identical precursor-product ion transitions, and (iii) reproduction of the selected precursor ionproduct ion relative abundances to within 10% of absolute for each target analyte relative to a standard of that analyte run under the same experimental conditions. On the basis of these criteria, the presence of selenocystamine and selenomethionine in the examined urine samples is clearly confirmed. The total Se concentration in the original urine samples was estimated to be approximately 90 ppb, and the concentration of selenomethionine and selenocystamine in the collected fractions was estimated to be approximately 11 and 40 ppb, respectively. ACKNOWLEDGMENT The authors acknowledge financial support from the Research Corp., under Grant No. RA0267, and from a Georgia Tech/CDC Seed Grant. The authors also thank the Environmental Separations Branch at the Georgia Tech Research Institute (GTRI) for access to the SCIEX API III mass spectrometer. Received for review January 3, 2001. Accepted April 8, 2001. AC0100244
