Elsevier

Journal of Chromatography A

Volume 1216, Issue 6, 6 February 2009, Pages 956-966
Journal of Chromatography A

Comparison of electrospray ionization, atmospheric pressure chemical ionization and atmospheric pressure photoionization for determining estrogenic chemicals in water by liquid chromatography tandem mass spectrometry with chemical derivatizations

https://doi.org/10.1016/j.chroma.2008.12.023Get rights and content

Abstract

This study compared the sensitivities and matrix effects of four ionization modes and four reversed-phase liquid chromatographic (LC) systems on analyzing estrone (E1), 17β-estradiol (E2), estriol (E3), 17α-ethinylestradiol (EE2), 4-nonylphenol (NP), 4-tert-octylphenol (OP), bisphenol A (BPA) and their derivatives of dansyl chloride or pentafluorobenzyl bromide (PFBBr) in water matrixes using a triple-quadrupole mass spectrometer with selected reaction monitoring (SRM). The four probes were electrospray ionization (ESI), atmospheric pressure chemical ionization (APCI), atmospheric pressure photoionization (APPI) and APCI/APPI; the four LC systems were ultra-performance liquid chromatography (UPLC) with or without post-column split, a mixed-mode column and two-dimensional LC (2D-LC). Dansylated compounds with ESI at UPLC condition had the most intense signals and less matrix effects of the various combinations of ionization and LC systems. The on-column limits of detection (LODs) of dansylated estrogens by SRM were 0.05–0.20 pg, and the LODs in sewage treatment plant effluent and in river water were 0.23–0.52 and 0.56–0.91 ng/L, respectively. The LODs using selected ion monitoring (SIM) reached low ng/L levels in real samples and measured concentrations were comparable with those of SRM.

Introduction

Feminizing contaminants of steroid estrogens, detergent degradates and plasticizers have caused a worldwide concern. They may influence the ecosystem at trace levels and affect human health through their contamination of drinking water. Natural estrogens 17β-estradiol (E2) and its synthetic analogue 17α-ethinylestradiol (EE2), an ingredient in oral contraceptives, are the most estrogenic. Moreover, their major metabolites, estrone (E1) and estriol (E3), are still bioactive. These steroid estrogens enter the water environment via the urine of humans and animals in the form of hydrophilic glucuronide and sulfate conjugates [1], which are biologically inactivated [2]. However, they are likely to be deconjugated in sewage treatment systems and converted to estrogenically active free forms [3]. 4-nonylphenol (NP), 4-tert-octylphenol (OP) and bisphenol A (BPA), which are all xenoestrogens, can affect normal endocrine functions. Although they are less potent, they are usually found in much higher concentrations in water (ng/L–μg/L) [3], [4], [5], [6]. These xenoestrogens are released into the water environment from daily usage of non-ionic surfactants and plasticizers.

Atmospheric pressure photoionization (APPI) is an emerging source, which is capable of ionizing nonpolar compounds and is possibly less susceptible to matrix effects. In addition, dual-source ionization (e.g. atmospheric pressure chemical ionization (APCI)/APPI combo in this study) expands the range of compounds that can be simultaneously analyzed. Although most studies determined feminizing chemicals with electrospray ionization (ESI) coupled with LC/MS(/MS) [7], [8], [9], the suitability of APCI and APPI deserve further exploration.

Matrix effect, which co-eluting components from the matrix or the mobile phase may enhance or suppress signals, is an important issue in using LC/MS/MS. Selective extraction, additional clean-up, efficient LC separation or change of mobile phase compositions may reduce matrix effects [10]. Furthermore, while the use of suitable internal standards (e.g. isotope-labeled chemicals) may correct signal irreproducibility, this approach will not be able to overcome the loss in sensitivity caused by matrix effects. Some studies utilized direct online extraction or post-column split to minimize matrix effects and simplify the sample preparation. A novel column developed on September 2006 combines both size exclusion and reverse-phase chemistry to separate small molecules from complex matrix [11]; to the best of our knowledge, one study has analyzed drugs in bovine serum using the mixed-mode column [11]. A restricted access material (RAM) pre-column, with a similar separation mechanism to the mixed-mode column, has also been applied on analyzing food, biological and environmental samples [12], [13]. However, the RAM pre-column is an alkyl-diol silica (ADS) column and provides little chromatographic separation for small molecules; it requires a column switch to connect it with an additional analytical column for chromatographic separation (a two-dimensional LC, 2D-LC). In addition, a post-column split delivers only a portion of LC flow into the MS, which may substantially decrease matrix effects, especially when a flow rate into ESI interface was decreased to nanoflow of 0.1 μL/min [14], [15]. This nanosplit requires special nanospray probes, which is not amenable to a conventional ESI interface, whose flow rates can be only as low as 20–50 μL/min. Reports on the mixed-mode column, 2D-LC and post-column split are very limited in environmental analysis and so little is known about their ability to reduce matrix effects.

Recently, there has been an increase in the number of studies using ultra-performance liquid chromatography (UPLC) combined with MS/MS. UPLC takes advantage of smaller packing particles (<2.0 μm) that enable high flow rates for fast chromatography without sacrificing separation efficiency, and signal-to-noise (S/N) ratios of analytes are increased because of sharp peaks. However, to best of our knowledge none have used UPLC/MS/MS to study estrogenic compounds in water.

Steroid estrogens and phenolic xenoestrogens are weak acids and their ionization on ESI and APCI are not very efficient compared with other more polar chemicals. Chemical derivatization can add on moieties improving ionization and enhance signals. For example, dansyl chloride or pentafluorobenzyl bromide (PFBBr) can react with phenolic groups, significantly improving sensitivity [16], [17], [18]. By adding the dansyl moiety with ESI interface, signal intensity may be increased as much as three orders of magnitude [16], [19], [20]. This technique has been also found to improve the sensitivity in APCI interface when used to measure steroid estrogens [21]. To date no dansyl derivatives have been analyzed with APPI interface. PFBBr derivatives can capture soft electrons in APCI, resulting in unstable metastic ions, and cause subsequent dissociation to generate negative ions through the loss of pentafluorobenzyl radical (electron-capture atmospheric pressure negative ionization, EC-APNI). For estrone, the use of PFBBr derivatives in EC-APNI can enhance efficiency of ionization as much as 25 times that of APCI alone [18]. This method has been also used in APPI with a high toluene dopant flow rate (e.g. 200 μL/min or higher) and was found to be able to detect as little as 0.17 pg of 2,4-dinitrophenol [22], whereas for PFBBr-derivatized estrone, signal enhancement (1.4–9.8 times) was less than that using EC-APNI [18]. Our group previously reported that dansylated estrogens with ESI interface provided better signal intensities than that PFBBr derivatives with EC-APNI, but obvious signal suppression was encountered with ESI when analyzing complex matrixes such as river water and effluents from sewage treatment plants [23].

In this study, we investigated signal intensity and matrix effects on various chromatographic systems (UPLC with or without flow split, mixed-mode column, 2D-LC) and several ionization modes (ESI+, ESI−, APCI+, APCI−, APPI+, APPI−, APCI/APPI+, APCI/APPI−) for both estrogenic compounds and their derivatives of dansyl chlorine and PFBBr. In addition, the study is unique in that it first optimized the operation conditions specific for each ionization methods, including those for LC columns, mobile phase flow rates and compositions. Previous studies usually compared the performance of different ionization sources under only one analytical column kept at a constant solvent flow rate, isocratic chromatography, the same injection volume, or flow injection analysis alone [24], [25]. However, the conclusions based on the results using non-optimized parameters of various ionization methods could be controversial. The main purpose of the study was to find out the best combination of a chromatographic system and an ionization method with satisfactory sensitivity using low volumes of water samples or single quadrupole MS. The final method was validated using river water and effluents from a sewage treatment plant (STP).

Section snippets

Chemicals and reagents

Estrone, 17β-estradiol, estriol, 17α-ethinylestradiol, 4-tert-octylphenol, bisphenol A, and bisphenol A-d16 (as a recovery standard) were obtained from Sigma/Aldrich (Saint Louis, MO, USA; purity > 98%). The technical mixture of nonylphenol was supplied by Riedel-de Haën (Seelze, Germany; purity > 94%). 2,4,16,16-2D4-estrone, 2,4,16,16-2D4-17β-estradiol, 2,4,17-2D3-16α-hydroxy-17β-estradiol, 2,4,16,16-2D4-17α-ethinylestradiol and 4-n-Octyl-d17-phenol were bought from C/D/N Isotopes (Pointe-Claire,

Effects of dopant, mobile phase flow rates and compositions on APPI sensitivity

We found that excess dopant flow may or may not improve analyte signals, and optimal dopant portions were compound-dependent. Different amount of the toluene dopant were tested on a Waters BEH C18 column, ranging from 5% to 25% of a constant mobile phase flow rate at 500 μL/min. The mobile phase compositions for native and dansylated compounds were Milli-Q water/methol and 10 mM formic acid/acetonitrile, respectively, which were the optimal combinations for their separation on the column. Based

Conclusions

In this study, we present a quantitative method for the analysis of seven estrogenic compounds with dansyl derivatization in both SIM and SRM. With the improvement in sensitivity using UPLC and chemical derivatization method, environmental levels of these chemicals can be determined using a single MS instead of the more expensive tandem-MS. The instrumental throughput was significantly increased, with a run in 3.2 min plus 1-min re-equilibrium time. We exhaustively investigated the performance

Acknowledgement

This work is supported by the National Science Council, Taiwan (NSC 96-2314-B-002-101-MY2).

References (35)

  • A.C. Johnson et al.

    Sci. Total Environ.

    (2000)
  • B. Stachel et al.

    Environ. Pollut.

    (2003)
  • U. Bolz et al.

    Environ. Pollut.

    (2001)
  • S. Rodriguez-Mozaz et al.

    J. Chromatogr. A

    (2004)
  • L. Salste et al.

    Sci. Total Environ.

    (2007)
  • C.Y. Chen et al.

    Sci. Total Environ.

    (2007)
  • T. Benijts et al.

    J. Chromatogr. A

    (2004)
  • J. Chico et al.

    J. Chromatogr. A

    (2008)
  • C.L. Andrews et al.

    J. Chromatogr. A

    (2004)
  • A. Salvador et al.

    J. Chromatogr. A

    (2007)
  • F. Qin et al.

    Anal. Chim. Acta

    (2008)
  • R. Mohamed et al.

    J. Chromatogr. A

    (2007)
  • S.S. Cai et al.

    J. Chromatogr. A

    (2007)
  • L.P. Penzes et al.

    J. Chromatogr.

    (1970)
  • D.B. Robb et al.

    J. Am. Soc. Mass Spectrom.

    (2005)
  • T.J. Kauppila et al.

    J. Am. Soc. Mass Spectrom.

    (2005)
  • D.B. Robb et al.

    J. Am. Soc. Mass Spectrom.

    (2006)
  • Cited by (0)

    View full text