Dinah Schick, Claudia Oellig
Institute of Food Chemistry, University of Hohenheim, Stuttgart, Germany
Nonylphenols (NP) are ubiquitous substances that have been detected in highly diverse foodstuff and that inter alia show estrogenic activity.1,2 NP can presumably end up in the aquatic environment as a result of breakdown of nonylphenol ethoxylates (NPE), which are used as non-ionic surfactants.3 According to the water framework directive regarding water quality of the European Union, NP are classified as priority hazardous substances, and thus “present a significant risk to or via the aquatic environment”.4–7 Different isomers of NP are collectively classified as one group, and the maximum acceptable concentration in inland and other surface waters is stated at 2 µg/L.
Because of the high risk emanating from NP, and due to existing regulations, the determination of their total concentration is very important. For this purpose, planar solid phase extraction (pSPE)8 is perfectly suited; and the estrogenic properties of NP can be used for their selective detection by planar yeast estrogen screen (pYES).9 Thus, the combined approach pSPE– pYES was developed.10
The determination of NP as a group is realized by pSPE - a method based on high-performance thin- layer chromatography (HPTLC) - that enables the analysis of matrix-rich extracts on planar thin-layers by separating matrix compounds, while simultaneously focusing the analytes of interest in a common target zone. The highly selective detection and quantitation of NP is eventually performed by means of pYES. As a screening tool for estrogen active compounds (EAC), pYES - also based on HPTLC - uses genetically modified yeast cells and water-wettable reversed phase HPTLC (RP-18 W) plates. The human estrogen receptor and a reporter gene encoding for β-galactosidase, respectively, are integrated in the yeast cells, leading to the production of the enzyme in the presence of estrogenic substances.11,12 The enzyme produced subsequently cleaves the suitable substrate resorufin- β-D-galactopyranoside, releasing orange fluorescing resorufin as a positive signal of estrogenicity.
Figure 1. Workflow of planar solid phase extraction–planar yeast estrogen screen (pSPE–pYES). After application and pSPE, pYES was performed on the same plate by means of yeast and substrate incubations and subsequent detection of released orange fluorescing resorufin.*
The combination of pSPE and pYES on HPTLC RP-18 W plates was designed for the detection of EAC as a group, and was successfully applied to screen for estrogen active NP in surface waters. Prior to the analysis of environmental samples, the response of the pSPE–pYES for the detection of different commercially available NP mixtures was investigated. Additionally, its suitability for the analysis of NP aside from other possible EAC was proven.
By pSPE–pYES, NP were successfully separated from other EAC and simultaneously focused in a common zone. An example is indicated for the synthetic hormone EE2 and the natural hormone E2 (Figure 2). The experiments revealed similar responses of different purchased NP, except for NP3 that has a linear nonyl side chain. The NP with the highest response, NP4, that has a branched side chain, was used as the representative reference substance for future experiments.
Figure 2.Track images of six different nonylphenols (NP, each 200 ng/zone) and a track with 17α-Ethinylestradiol (EE2, 200 pg/ zone) and 17β-Estradiol (E2, 200 pg/zone) under UV 254 nm illumination after pSPE–pYES. NP1, NP2, NP4, NP5 and NP6 are technical mixtures of NP; NP3 is 4-n-NP (see Table 1 for conditions).
Sensitivity of pSPE–pYES and the recovery rate of the entire method (extraction of 200 mL water with 20 mL of dichloromethane by stirring for 10 min at 1000 rpm, separating and removing of the organic phase, and dissolving of the residue in 200 µL of ethanol)10 were determined to evaluate the performance of the screening. Limits of detection and quantitation (LOD and LOQ) were 14 ±4 and 26 ±4 ng NP/zone (each n=8), corresponding to ~1 µg/L LOD taking the extraction and application volume into account.10 Recovery of NP was 95 ±17% (n=12, n=4 extracts on n=3 days, Table 2).10
Screening of surface waters for NP was performed after extraction of samples by application of the extracts and standard solutions on HPTLC plates, followed by a quick focusing step, whereupon pSPE–pYES was performed. Water samples (n=7) were taken from different sites such as lakes, ponds, and streams, and were each analyzed as native samples and after spiking with NP at a concentration of 2 µg/L. Example images of the single steps of pSPE–pYES for extracts of pond water are shown in Figure 3. Estrogen active NP were not detected in any of the investigated water samples by the developed screening; however, the analysis of the spiked samples showed the applicability of the screening for matrix-containing environmental samples. The applied extraction method was also shown to be suitable as the recovery rate from spiked environmental samples (95 ±17%, n=7) was very similar to the recovery rate determined from spiked ultrapure water.10
Figure 3. Track images of extracts of a native and spiked (2 μg NP4/L) sample of surface water from a pond under UV 366 nm illumination after (A) application, (B) focusing step, (C) 1st development and (D) 2nd development, and (E) under UV 254 nm illumination after pYES. (F) track images after pSPE and pYES, respectively, of the same extracts and a blank extract and the respective 3D densitogram of the fluorescence scan at 550/>580 nm. (Partly modified and reprinted by permission from Springer*).
The advantage of pSPE—pYES was especially evident after analysis of environmental samples. Since no further cleanup of the extracts was necessary, all matrix compounds were applied onto the HPTLC plate, partly visible by native blue and red fluorescence (Figure 3). Despite the fact that the complete matrix was subjected to the analysis, estrogen active NP were clearly detectable, quantifiable, and differentiable from co-extracted substances.10 Since pSPE is based on HPTLC, it serves as planar cleanup, separating substances of interest from matrix compounds while the analytes are simultaneously focused in a common target zone.8 The high selectivity of the applied pYES is caused by the screening itself. Positive signals, shown by orange fluorescence of released resorufin, are based only on the activity of the yeasts and the reporter gene after estrogenic compounds have been bound by the receptor. This is why false positive signals or interferences are not to be expected.9,10
By combining pSPE and pYES on HPTLC RP-18 W plates, a meaningful and straightforward approach was developed for the rapid screening and detection of estrogen active NP in surface waters. By use of pSPE, complex sample treatments are redundant and matrix-rich extracts can directly be applied to analysis. At the same time, pSPE focuses the analytes of interest in a common zone, enabling the detection and quantitation as a whole. By combination with pYES, i.e., the detection of the analytes on the basis of their estrogenicity, pSPE–pYES represents a highly selective and innovative approach for the analysis of EAC.
* Reprinted by permission from Springer Nature Customer Service Centre GmbH: Springer, Schick, D.; Oellig, C., Screening for estrogen active nonylphenols in surface waters by planar solid phase extraction–planar yeast estrogen screen, Analytical and Bioanalytical Chemistry. 2019, 411, 6767-6775. (https://doi.org/10.1007/s00216-019-02053-0)