Anal. Chem. 2005, 77, 7032-7038
Analysis of Steroid Conjugates in Sewage Influent and Effluent by Liquid Chromatography-Tandem Mass Spectrometry Sharanya Reddy,† Charles R. Iden,‡ and Bruce J. Brownawell*,†
Marine Science Research Center and Department of Pharmacological Sciences, Stony Brook University, Stony Brook, New York 11794-5000
Environmental endocrine disruptors such as estrone (E1) and β-estradiol (E2) are excreted in human urine primarily as water-soluble glucuronides and sulfates that can dissociate in wastewater treatment systems to the more active free estrogens. Measurement of the distribution and fate of the steroid conjugates and the corresponding free estrogens in treatment plants and receiving waters is critical for understanding the reproductive and developmental effects of these substances on aquatic organisms. A sensitive method to measure steroid estrogen conjugates in matrix-rich sewage influents and effluents (method detection limits ranged from 0.04 to 0.28 ng/L) has been developed using HPLC tandem mass spectrometry with electrospray ionization. The method employs extensive sample purification by selective extraction from an Oasis HLB solid-phase cartridge followed by separation by anion exchange chromatography. This purification scheme, combined with a stable isotope dilution approach, was used to overcome problems of matrix suppression of ionization and permitted selective and sensitive detection of six target conjugates of E1 and E2. Accurate quantitation was highly dependent on the method of sample preservation. Acidification of each sample (pH 2.0) was effective in preventing enzymatic or chemical decomposition of steroid conjugates in all sample types, whereas glucuronide conjugates were hydrolyzed in the presence of mercury and formalin preservatives. Measured concentrations of steroid sulfates in the influent to a sewage treatment plant were ∼100 times greater than that of the respective steroid glucuronides, suggesting that the preponderance of glucuronides had dissociated prior to reaching the treatment plant. A small percentage of the steroid sulfates persisted through biological treatment of sewage and was measured in the effluent. Steroid conjugates that survive decomposition or bypass biological treatment of municipal wastewater are released into surface waters and may serve as a source of free steroids.
* Corresponding author: E-mail:
[email protected]. Fax: 631-632-8820. † Marine Science Research Center. ‡ Department of Pharmacological Sciences.
7032 Analytical Chemistry, Vol. 77, No. 21, November 1, 2005
The presence of endocrine-disrupting chemicals in the aquatic environment is generating worldwide concern because these chemicals may cause feminization of male and sexually immature fish1,2 and interfere with reproduction and development of aquatic organisms. Among known endocrine disruptors, the naturally occurring estrogenic hormones, estrone (E1) and β-estradiol (E2), secreted by humans and animals and the synthetically prepared contraceptive, ethynylestradiol (EE2), used by women show the highest degree of estrogenic activity in aquatic environments.3,4 Steroid hormones, including E1 and E2, are excreted in urine of humans and animals as the more hydrophilic glucuronides and sulfates;5 however, these steroid conjugates do not have affinity to estrogen receptors. Evidence suggests that once the conjugated estrogens reach sewage treatment plants, they undergo chemical or enzymatic dissociation in bacterial sludge and re-form active estrogens.6,7 Thus, it is important to determine the fate and distribution of the steroid conjugates in the environment since they are potential sources of active estrogens due to dissociation in sewage treatment plants or as a result of the input of untreated wastewater directly into surface waters. Several analytical methods have been developed to identify and quantify E1, E2, and EE2 in sewage-impacted rivers, estuaries, and marine environments using both chemical techniques such as gas chromatography-mass spectrometry (GC/MS),8-10 highperformance liquid chromatography-mass spectrometry (HPLCMS),11,12 and immunoassays.13,14 All of these methods involve (1) Routledge, E. J.; Sheahan, D.; Desbrow, C.; Brighty, G. C.; Waldock, M.; Sumpter, J. P. Environ. Sci. Technol. 1998, 32, 1559-1565. (2) Tyler, C. R.; Jobling, S.; Sumpter, J. P. Crit. Rev. Toxicol. 1998, 28, 319361. (3) Desbrow, C.; Routledge, E. J.; Brighty, J. P.; Sumptor, J. P.; Waldock, M. Environ. Sci. Technol. 1998, 32, 1549-1559. (4) Fang, H.; Tong, W.; Perkins, R.; Soto, A. M.; Prechtl, N. V.; Sheehan, D. M. Environ. Health Perspect. 2000, 108, 723-739. (5) Shackleton R. H. L. J. Chromatogr. 1986, 91-156. (6) Ternes, T. A.; Stumpf, M.; Mueller, J.; Haberer, K.; Wilken, R.-D.; Servos, M. Sci. Total. Environ. 1999, 225, 81-90. (7) Ternes, T. A.; Mueller, P. K. Sci. Total Environ. 1999, 225, 91-99. (8) Daughton, C. J.; Ternes, T. A. Environ. Health Perspect. 1999, 107, 907938. (9) Alda, M. J. L.; Diaz-Cruz, S.; Petrovic, M.; Barcelo, D. J. Chromatogr. 2003, 1000, 503-526. (10) Vanderford, B. J.; Pearson, R. A.; Rexing, D. J.; Snyder, S. A. Anal. Chem. 2003, 75, 6265-6274. (11) Alda, M. J.; Barcelo, D. J. Anal. Chem. 2001, 371, 437-447. (12) Ferguson, P. L.; Iden, C. R.; McElroy, A. E.; Brownawell, B. J. Anal. Chem. 2001, 73, 3890-3895. (13) Huang, C. H.; Sedlak, D. L. Environ. Toxicol. Chem. 2001, 20, 133-139. 10.1021/ac050699x CCC: $30.25
© 2005 American Chemical Society Published on Web 09/28/2005
extensive extraction and purification steps. Very few analytical methods have been developed to quantify steroid conjugates in wastewater samples. Conjugates cannot be analyzed directly by GC/MS techniques but must undergo enzymatic or chemical dissociation to the free estrogens prior to analysis.15,16 Investigators have measured steroid conjugates using GC/MS by studying the difference in concentrations of steroid hormones in samples before and after enzymatic hydrolysis.17 Trace-level analysis of estrogen conjugates by such an approach is complicated when the difference between free and bound estrogen is small, hydrolysis is not 100% efficient, or the enzymes do not act on all forms of the conjugates.5,18 Acid hydrolysis of steroid conjugates is also problematical, often exhibiting uncontrolled losses of the free steroids.18 To circumvent the issue of dissociation of steroid conjugates, recent methods measure concentrations of steroid conjugates using HPLC-MS.19,20 Analysis of steroid estrogen conjugates by HPLC-MS has primarily focused on applications with biological fluids rather than wastewater where the sample matrix may be more complex.19-21 There are, however, two methods reported by Gentili et al.22 and Isobe et al.,23 which use tandem MS-MS for detection of the steroid conjugates in a multiple reaction monitoring (MRM) experiment in wastewaters. Both of the methods by Gentili et al.22 (this method was applied in a detailed study by D'Ascenzo et al.24) and Isobe et al.23 show low detection limits in relatively clean river and lake water samples, respectively. However, in more matrix rich sewage samples, detection limits were reported to be much higher.22,24 In our attempt to apply the method of Isobe et al.23 to analysis of sewage treatment plant effluents, we found that coeluting matrix resulted in greater than 80-90% reduction in electrospray ionization. While the effects of sample matrix are going to depend on sample origin and possibly on the design of the electrospray source, these results highlight the need for minimizing the effects of matrix on the analysis of steroid estrogen conjugates in wastewater samples, when the need for very low detection is required. In this work, the focus has been on developing methods that minimize isobaric interferences and suppression of ionization that complicate analysis in HPLC-ESI-MS-MS analysis of estrogen conjugates. With extensive purification of the samples, we were able to process and analyze larger sample volumes (∼30-fold greater than the method by Gentili et al.22 when post-HPLC column splitting is considered) without observing loss of sensitivity due to ion suppression. Also addressed in this study was the ability (14) Takigama, H.; Taniguchi, N.; Matsuda, T.; Yamada, M.; Shimizu, Y.; Matsui, S. Water Sci. Technol. 2000, 42, 45-51. (15) Masse’, R.; Ayotte, C.; Dugal, R. J. Chromatogr. Biomed. Appl. 1989, 489, 23-30. (16) Schanzer, W.; Donike, M. Anal. Chim. Acta 1993, 275, 23-28. (17) Belfroid, A. C.; Horst, Van der A.; Vethaak, A. D.; Schafer, A. J.; Rijs, G. B. J.; Wegener, J.; Cofino, W. P. Sci. Total Environ. 1999, 225, 101-108. (18) Axelson, M.; Sahlberg, B. L.; Sjovall, J. J Chromatogr. 1981, 224, 355-370. (19) Bowers, L. D.; Sanaullah. J. Chromatogr. 1996, 687, 61-68. (20) Zhang, H.; Henion, J. Anal. Chem. 1999, 71, 3955-3964. (21) Kuuranne, T.; Kotiaho, T.; Pederson-Bjergaard, S.; Rasmussen, K. E.; Leinonen, A.; Westwood, S.; Kostiainen, R. J. Mass Spectrom. 2003, 38, 16-26. (22) Gentili, A.; Perret, D.; Marchese, S.; Mastropasqua, R.; Curini, R.; Corcio, A. D. Chromatographia 2002, 56, 25-32. (23) Isobe, T.; Shiraishi, H.; Yasuda, M.; Shinoda, A.; Suzuki, H.; Morita, M. J. Chromatogr. 2003, 984, 195-202. (24) D’Ascenzo, G. D.; Gentili, A.; Mancini, R.; Mastropasqua, M.; Nazzari, M.; Samperi, R. Sci. Total. Environ. 2003, 302, 199-209.
of various sample preservation methods to prevent hydrolysis of estrogen conjugates during short-term sample storage. EXPERIMENTAL SECTION Materials. The glucuronide and sulfate conjugates of estrone, i.e., estrone-3-sulfate (E1-3S) and estrone-3-glucuronide (E1-3G) and β-estradiol, i.e., β-estradiol-3-glucuronide (E2-3G), β-estradiol17-glucuronide (E2-17G), β-estradiol-3-sulfate (E2-3S), and β-estradiol-17-sulfate (E2-17S), were purchased from Steroids Inc. (Newport, RI). The stable isotope surrogate standards for the sulfates of E1 (E1-3S-d4) and E2 (E2-3S-d4) were obtained from C/D/N Isotopes, Inc. (Pointe-Claire, PQ, Canada). Deuterated isotope surrogates of steroid glucuronides were synthesized enzymatically. Starting materials, estradiol-d4 (150 µM) and estrone-d4 (150 µM), were each incubated at 37 °C in 50 mM Tris (pH 7.5) in the presence of 2 mM uridine diphosphoglucuronic acid, magnesium chloride (10 mM), and human UGT1A1 supersomes (1 mg/mL, BD Biosciences, Woburn, MA). After incubation for 30 min, the reaction mixture was quenched with 50 µL of 94% acetonitrile/6% acetic acid and centrifuged for 3 min. The supernatant was diluted to 5 mL with water and loaded on an Oasis HLB cartridge (0.5 g, Waters, Milford, MA) preconditioned with 10 mL of methanol followed by 10 mL of MilliQ (MQ) water. The cartridge was washed first with ethyl acetate (8 mL) to remove any unreacted E1-d4 or E2-d4, then with 5 mL of 60% methanol containing 2% acetic acid, followed by 5 mL of 20% methanol containing 2% ammonium hydroxide. The E2-3G-d4 and E1-3G-d4 were eluted from each of the cartridges using 8 mL of 70% methanol containing 2% ammonium hydroxide. The eluant was dried under vacuum in a Savant SpeedVac and reconstituted in 100% methanol. The concentration of the deuterated glucuronide surrogate standard of E1 and E2 was estimated by HPLC-TOFMS by comparison with the respective undeuterated standards, E1-3G and E2-3G. Product recovery was 3 and 14% for E1-3Gd4 and E2-3G-d4 respectively; purity of the standards was 97% for E1-3G-d4 and 98% for E2-3G-d4. Radioactive E2-3G was synthesized to assist with method development using the same techniques by incubating 3H-estradiol (NEN Brand Radiochemicals, Perkin-Elmer, Boston, MA). Radioactivity was counted using a Tri-Carb 2100TR liquid scintillation analyzer (∆ Packard, Canberra Co., Meriden, CT). Methods. Wastewater Collection. Sewage influent (0.5 L) and effluent samples (1 L) were collected from the Stony Brook treatment plant located on the campus of Stony Brook University. This plant has a flow capacity of just over 1 million gal/day, serves a population of ∼15 000, and employs activated sludge secondary biological treatment technology. The samples were collected in baked (450 °C) amber glass bottles, immediately preserved by addition of sulfuric acid to pH of 2.0, and stored at 4 °C prior to analysis (up to 24 h). The addition of sulfuric acid did not cause hydrolysis or other transformations of conjugates over 7 days in experiments conducted at 4 °C over the pH range of 1.1-2.5. Experiments were conducted with radiolabeled E2-3G (0.018 µCi, specific activity 100 mCi/µmol) to determine the effectiveness of various approaches for preserving steroid estrogen conjugates. Preliminary studies had shown rapid loss of isotopically labeled conjugates over a couple hour period between collection, sample filtrations, and extractions. Tritiated E2-3G (0.5 ng/L) was added into sewage effluents and influents (100 mL) in the presence of Analytical Chemistry, Vol. 77, No. 21, November 1, 2005
7033
mercury chloride (50 mg/L) or formaldehyde (0.4%), or they were acidified to pH 2.0 with sulfuric acid. The spiked samples were left at room temperature from 2 to 24 h before loading on Oasis HLB cartridges and processed using an solid-phase extraction (SPE) method below to separate free estrogens from conjugated forms. Solid-Phase Extraction and Cleanup. The sewage influent and effluent samples were spiked with surrogate standards (5 ng each of E1-3S-d4, E2-3S-d4, E1-3G-d4, E2-3G-d4, E1-d4, and E2-d4) and were vacuum filtered using glass fiber filters (GF/C filters, 1.2 µm, Whatman). The samples were then loaded on Oasis HLB cartridges preconditioned with 10 mL of methanol followed by 10 mL of MQ water. After loading the sample, the cartridge was washed with 10 mL of MQ water followed by drying under vacuum for 1 h. The Oasis HLB sorbent (polydivinylbenzene-co-N-vinylpyrrolidine) binds both hydrophilic and lipophilic compounds. The more lipophilic E1 and E2 and their corresponding surrogates were eluted from the column with ethyl acetate (8 mL); the fraction was dried under nitrogen and analyzed for E1 and E2 using the method described below. The more hydrophilic steroid conjugates remained on the cartridge and were eluted after additional wash steps. Basic interferences were reduced first by washing with 30% methanol/water containing 2% acetic acid (5 mL) followed by 60% methanol/water containing 2% acetic acid (8 mL). Acidic interferences were removed by washing with 30% methanol/water containing 2% ammonium hydroxide (8 mL). The steroid conjugates were then eluted with 75% methanol/water containing 2% ammonium hydroxide (8 mL), and the fractions were dried under vacuum in a Savant SpeedVac. Anion Exchange Cleanup. Estrogen conjugate fractions from the SPE were reconstituted (50 µL) in 0.05 mM sodium monohydrogenphosphate (pH 7.8, adjusted with 10% HCl), prior to loading over a weak anion exchange column (DEAE, Nucleogen 60-7, 125 × 4.6 mm, Macherey-Nagel) and chromatographic separation using Shimadzu HPLC pumps (LC600, Shimadzu, Kyoto, Japan). Initial experiments with solvents to elute conjugates from the anion exchanger showed that higher salt concentrations were needed to elute the analytes and recoveries were poor if the salts did not include phosphates. Fractions were collected from the column using a salt gradient (solvent A, 0.05 mM sodium monohydrogenphosphate containing 10% acetonitrile, pH 7.8; solvent B, 100 mM sodium monohydrogenphosphate containing 500 mM NaCl and 10% acetonitrile, pH 7.0). Gradient conditions involved initial flushing of the column for 3 min with 100% solvent A at 1 mL/min followed by 0-30% increase in solvent B over 22 min, 30-70% increase in solvent B over 2 min, maintain 70% solvent B for 5 min, and followed by equilibration of the column back to solvent A. Steroid conjugates were detected at 284 nm with an online Shimadzu UV-visible spectrophotometer (Shimadzu, model SPD-6AV). All steroid conjugates eluted between 10 and 16.4 min, and a single fraction corresponding to this time window was collected. Fractions were dried under vacuum and desalted on C18 (E-C) cartridges (200 mg, Isolute, Mid-Glamorgan, U.K.) preconditioned with 10 mL of methylene chloride, followed by 10 mL of methanol and 10 mL of MQ water. The dried fractions were reconstituted in 5 mL of MQ water and loaded on the preconditioned cartridges. The cartridges were washed with MQ water (10 mL), and the samples were eluted with 80% methanol/ 7034
Analytical Chemistry, Vol. 77, No. 21, November 1, 2005
water (5 mL). The desalted samples were dried under vacuum prior to analysis by HPLC-MS-MS. HPLC-MS-MS Analysis for Steroid Conjugates. Concentrations of estrogen conjugates were determined in sewage effluent and influent samples using HPLC-MS-MS. The conjugates were separated on a Zorbax Extend-C18 column (3.5-µm particle size, 100 mm × 2.1 mm i.d., Agilent Technologies) by HPLC (HewlettPackard, HP1100 binary pump and autosampler, Foster City, CA). A gradient separation was achieved using two solvents: solvent A, MQ water containing 13.5 mM ammonium hydroxide; solvent B, acetonitrile containing 13.5 mM ammonium hydroxide. Gradient conditions were initiated with 5% solvent B (maintained for 2 min) followed by a linear gradient from 5 to 70% solvent B in 20 min. The conditions were maintained at 70% B for 2 min before the column was reequilibrated to starting conditions. The desalted samples were reconstituted in solvent A (final volume, 100 µL) containing internal standard Equilin sulfate-d4 (1.02 ng), and 25 µL of each sample was injected. The mass spectrometer was a Quattro LC (Micromass, Manchester, U.K.) equipped with a Z-Spray ESI source. The mass spectrometer was operated in the negative electrospray ionization mode using multiple reaction monitoring (MRM). The conditions for detection by the mass spectrometer were as follows: desolvation gas flow, 200 L/h; capillary voltage, 2.71 kV; multiplier voltage, 650 V. To obtain high sensitivity during the MRM analysis, each chromatographic run was divided into two acquisition time periods. In the first period between 7 and 13.6 min, intensities of ions for E1-3G, E2-3G, E2-17G, E1-3G-d4, E2-3G-d4, and E217S were monitored, while in the second acquisition period between 13.6 and 18.0 min, intensities of ions for E1-3S, E2-3S, E1-3S-d4, and E2-3S-d4 were detected. Five-point calibration curves were made for each of the steroid conjugates within the linear range of the instrument (0.3-65.0 ng/mL). HPLC-Time-of-Flight-Mass Spectrometry Free Steroid Analysis. The dried ethyl acetate fraction obtained from the SPE extraction and purification procedure was reconstituted in 25 mL of 1% methanol. Extracts were purified on a synthetic immunoaffinity column following the procedure of Ferguson et al.12 The eluents were evaporated to dryness, and the sample was reconstituted in 200 µL of 25% acetonitrile in MQ water. The equilin-d4 internal standard (6 ng) was added to each sample prior to analysis by LC-MS. The steroids were separated on a Betasil C18 column (3-µm particle size, 15 cm × 2.1 mm i.d.; Keystone Scientific Inc., Bellefonte, PA) using a Waters 2695 LC (Waters, Milford, MA) with HPLC conditions described previously.12 The steroids were detected as negative ions using a Micromass LCT TOF mass spectrometer, equipped with a 4.6-GHz time-to-digital converter and a Z-spray electrospray source. The electrospray conditions were optimized with a capillary voltage of -2200 V, sample cone at -60 V, and source temperature at 120 °C. The nitrogen desolvation gas flow rate was 406 L/h at 200 °C. The multichannel plate detector was operated at 2750 V, and the instrument was externally calibrated using polyalanine. Four-point calibration curves were made for the steroids within the linear range of the instrument (0.5-10 ng/mL, r2 > 0.999). The concentrations of E1 and E2 in the samples were calculated relative to the surrogate standards. Both E1 and E2 were internally mass calibrated with respect to their corresponding surrogate standards using the
Table 1. Sample Preservation Conditions % radioactivity
sample
preservative
in free steroid fraction
influent (2 h) influent (2, 24 h) influent (2, 24 h) influent (2 h) water effluent (24 h) effluent (4 h)
unpreserved HgCl2 pH 2.0 formalin pH 2.0 pH 2.0 HgCl2
80 8.8, 8.0 6.2, 7.0 30 6.8 6.0 65
in steroid conjugate fraction 16 90, 87 84, 86 65 84 80 35
Micromass all file accurate mass measure (AFAMM) software. RESULTS AND DISCUSSION Sample Preservation. Since steroid conjugates are labile and can be degraded by bacterial conversion to free steroids and other products,7 conditions for the preservation of steroid glucuronides in water samples were determined. In the absence of any preservative, 80% of the spiked radioactive E2-3G in influent sample was converted to more lipophilic products (assumed here to be E2 or E1) within 2 h (Table 1). In the presence of formalin, almost 30% of the E2-3G was converted to free steroids in the influent sample in 2 h. The results from control and selected preserved samples indicate a small portion of E2 3-G elutes (7%) in the free steroid fraction; thus, there is no evidence of chemically or biologically mediated hydrolytic conversion of E2-3G in the samples preserved at pH 2. In the effluent samples, radioactive E2-3G was preserved only at acid pH, not by the addition of mercury chloride (∼65% conversion in 4 h). It is unclear why mercury was apparently effective in preserving E2-3G in the influent but not the effluent sample. The inability of mercury to preserve glucuronide conjugates is consistent with the structure of glucuronidase enzymes that are likely responsible for the observed transformations. Mercury is known to inhibit enzyme activity by binding to sulfhydryl groups in cysteine or methionine residues often present at active sites.25 Experiments with β-glucuronidase indicated that enzyme activity was unaffected in the presence of 100-fold higher molar concentration of mercury chloride than used for sample preservation (data not shown), consistent with the fact that the active site of this enzyme contains a glutamate residue, not a sulfhydryl group.26 In the case of the effluent, there are indications that extracellular β-glucuronidase is produced by bacteria in wastewater during activated sludge treatment.7 The activity of this extracellular glucuronidase in the effluent is not inhibited by mercury chloride. There may be lower concentrations of extracellular glucuronidase or greater concentrations of inhibitors of the extracellular hydrolytic enzymes in the matrix-rich influent. It is likely the primary action of mercury is on the bacterial cells in the influent, where it may target mitochondrial respiration27,28 and inhibit metabolic (25) Vignes, M.; Guiramand, J.; Sassetti, I.; Recasens, M. Eur. J. Neurosci. 1993, 5, 327-334. (26) http://us.expasy.org/ Search in Swiss-Prot/TrEMBL for P05804. (27) Konisberg, M.; Lopez-Diazguerrero, N. E.; Bucio, L.; Gutierrez-Ruiz, M.-C. J. Appl. Toxicol. 2001, 21, 323-320. (28) Lund, B. O.; Miller, D. M.; Woods, J. S. Biochem. Pharmacol. 1991, 42 (Suppl), S181-187.
activity of the cells, preventing production of extracellular bacterial glucuronidase. From this work it was clear that steroid conjugates in sewage influent and effluent were best preserved by the presence of acid. SPE and Anion Exchange Cleanup. Samples underwent a rigorous cleanup prior to analysis by HPLC-MS-MS. The sorbent in the Oasis HLB cartridge bound both the relatively lipophilic steroids and the hydrophilic steroid conjugates as observed by Isobe et al.23 The free steroids were eluted in the ethyl acetate, and the steroid conjugates were extracted from the remaining material on the cartridges after additional selected wash steps. Basic compounds with a hydrophobic character were reduced using 60% methanol/2% acetic acid; relatively hydrophilic acidic compounds were removed using 30% methanol/2% ammonium hydroxide, leaving acidic compounds that were relatively more hydrophobic, including steroid conjugates. Recovery was determined in spiked controls of the conjugates (5 ng each) in acidified MQ water (pH 2.0) relative to the internal standard equilin sulfate-d4 and was estimated to be 81-93% for all the conjugates. The stepwise washes with methanol containing acid or base prior to elution of the steroid conjugates over the Oasis HLB cartridge helped to reduce isobaric interferences and also reduced ion suppression as measured by the external response of the internal standard equilin sulfate-d4 from 90 to 70%. However, additional purification of sample extracts by ion exchange chromatography proved necessary to further reduce extensive ion suppression of the internal standard. The ion suppression was reduced to 35% in the influent sample purified over SPE followed by anion exchange. This remaining ion suppression of the analytes observed in the samples could be accounted for by the method of isotope dilution. The effect of anion exchange purification also led to substantial reduction in isobaric interferences for all of the analytes and a resulting increase in signal-to-noise ratio and level of confidence in analyte confirmation. Prior work22,23 did not address in detail the issue of minimizing ion suppression or isobaric interference that can complicate ultratrace ESI-MS-MS analysis of steroid estrogens in any method. The lack of ion suppression observed by Gentili et al.22 may be partly influenced by the amount of sample processed (100 mL of influent and 250 mL of effluent) and analyzed by ESI-MS after postcolumn splitting. The other likely reasons for lack of ion suppression may be a difference in sample purification methods or a difference in ionization sources, since Gentili et al.22 used a PE Sciex API mass spectrometer that has a different ESI source design HPLC-MS-MS. Negative ESI analysis consistently produced more intense ion signals for steroid estrogen conjugates than positive ion analysis. Since electrospray ionization is largely dependent on the solvent conditions used during analysis, we investigated the negative ionization of steroid conjugates in the presence of different organic solvents and both acidic and basic pH. Steroid conjugates yielded a 2-3-fold increase in ion intensity when dissolved in acetonitrile/water compared to methanol/water mixtures, consistent with observations made previously.29 The difference in response may be due to differences in viscosity between the acetonitrile/water and methanol/water mixtures, (29) Benijts, T.; Dams, R.; Gunther, W.; Lambert, W.; Leenheer, A. D. Rapid Commun. Mass Spectrom. 2002, 16, 1358-1364.
Analytical Chemistry, Vol. 77, No. 21, November 1, 2005
7035
Table 2. MRM Settings for Steroid Conjugates compound
abbrev
estrone-3-sulfate d4 estrone-3-sulfate estradiol-3-sulfate estradiol-17-sulfate d4estradiol-3-sulfate estrone-3-glucuronide d4 estrone-3-glucuronide estradiol-3-glucuronide estradiol-17-glucuronide d4estradiol-3-glucuronide
E1-3S d4 E1-3S E2-3S E2-17S d4 E2-3S E1-3G d4 E1-3G E2-3G E2-17G d4 E2-3G
transition ion
which may result in faster creation of droplets with the required size for ion formation.29 As observed elsewhere,22 the addition of a strong base such as triethylamine (TEA) or ammonium hydroxide increased the molecular ion intensity of steroid glucuronides by almost 3-fold; however, base had little effect on the ionization of more acidic sulfate conjugates. NH4OH was chosen over TEA, since our source of TEA contained impurities that coeluted with the analytes during HPLC separation. A Zorbax Extend-C18 column was employed for HPLC separation, since the end-capping made the column more resistant to the base used in the mobile phase. The steroid estrogen conjugates were analyzed by tandem MS-MS in the negative ion MRM mode. The MRM transitions, cone voltages, and collision energies were optimized for each compound by infusing standard solutions into the mass spectrometer (Table 2). Single ion transitions were monitored for all the analytes which were characterizitic of the parent compounds except in the case of E2-17S, for which the sulfate ion was used for a product ion. The fragmentation of the molecular ion of the E2 and E1 glucuronides and their isotopically labeled surrogates resulted in a base peak corresponding to the neutral loss of the glucuronide moiety [M - H - Glu]- similar to observations in previous studies with anabolic steroids.30 In the case of steroid sulfates, daughter ion spectra were characterized by base peaks [M - H - 80]- (loss of SO3) when the sulfate group was attached to an aromatic ring (E2-3S and E1-3S). When the sulfate was bound to an alicyclic ring (E2-17S), the base peak was m/z 97 (HSO4-) consistent with observations by Zhang and Henion.20 It appears that a sulfate located on an aromatic ring (in this case, the 3-position) does not efficiently abstract a hydrogen from the 3-position and is unable to form an ion at m/z 97.31 We did not monitor for a second transition22 since, for most of the analytes, the second product was observed in very low abundance, and monitoring a second transition for confirmation would compromise the sensitivity of detection. However, in this work, an additional level of analyte confirmation was obtained using chromatographic retention of the deuterated standards. Figure 1 illustrates a standard mixture of conjugates (125 pg of each on column) analyzed by tandem MS-MS in the MRM mode. Figure 2 shows a representative chromatogram of the steroid conjugates and their corresponding deuterated standards in the influent sample purified over SPE and anion exchange column. (30) Bean, K. A.; Henion, J. D. J. Chromatogr., B 1997, 65-75. (31) Weildolf, L. O. G.; Lee, E. D.; Henion, J. D. Biomed. Mass Spectrom. 1998, 15, 283-289.
7036 Analytical Chemistry, Vol. 77, No. 21, November 1, 2005
E1 E1 E2 HSO4d4 E2 E1 d4 E1 E2 E2 d4 E2
MRM ions
cone (V)
collision (V)
349>269 353>273 351>271 351>97 355>275 445>269 449>273 447>271 447>271 451>275
42 42 42 41 55 46 46 53 53 53
34 34 34 36 25 36 36 30 30 30
Figure 1. Analysis of steroid conjugates by HPLC-MS-MS. The figure shows 0.125 ng of each of the steroid conjugates separated on a C18 column analyzed by HPLC tandem mass spectrometry. Equilin sulfate-d4 was spiked as internal standard (IS) into samples (10.2 ng/mL) prior to analysis by HPLC-MS-MS.
Method Validation. Calibration curves were obtained by analysis of 0.3-65 ng/mL of each analyte estimated as the relative response to the internal standard or the corresponding deuterated surrogate standards. Curves were linear with r2 values higher than 0.99. The instrument detection limits for the conjugates corresponded to sample concentrations between 0.03 and 0.08 ng/ L. Recoveries of the steroid conjugates were determined by spiking known amounts of each of the conjugates in acidified MQ water (pH 2-2.5, adjusted with sulfuric acid, n ) 5). The absolute recoveries of each of the conjugates were determined as a ratio to the internal standard, equilin sulfate-d4, added just prior to analysis of the sample by LC-MS (Table 3) and varied between 62 and 81% (standard deviation,
