Non-target estrogenic screening of 60 pesticides, six plant protection products, and tomato, grape, and wine samples by planar chromatography combined with the planar yeast estrogen screen bioassay

Abstract

For non-target residue analysis of xenoestrogens in food, sophisticated chromatographic-mass spectrometric techniques lack in biological effect detection. Various in vitro assays providing sum values encounter problems when opposing signals are present in a complex sample. Due to physicochemical signal reduction, cytotoxic or antagonistic effect responses, the resulting sum value is falsified. Instead, the demonstrated non-target estrogenic screening with an integrated planar chromatographic separation differentiated opposing signals, detected and prioritized important estrogenic compounds, and directly assigned tentatively the responsible compounds. Sixty pesticides were investigated, ten of which showed estrogenic effects. Exemplarily, half-maximal effective concentrations and 17β-estradiol equivalents were determined. Estrogenic pesticide responses were confirmed in six tested plant protection products. In food, such as tomato, grape, and wine, several compounds with an estrogenic effect were detected. It showed that rinsing with water was not sufficient to remove selected residues and illustrated that, though not usually performed for tomatoes, peeling would be more appropriate. Though not in the focus, reaction or breakdown products that are estrogenic were detected, underlining the great potential of non-target planar chromatographic bioassay screening for food safety and food control.

Keywords: Bioautography; Endocrine-disrupting compounds; Food safety; Planar yeast estrogen screen; Residue analysis.

Fig. 1

Pesticide screening: HPTLC–pYES–FLD bioautograms at 366 nm of pesticides (Table S1) showing MU-blue fluorescent estrogenic responses and corresponding biodensitograms at 366 nm/ > 400 nm for (a) fludioxonil (Flu; 0.06–1 µg/band), fenhexamid (Fen; 0.1–2 µg/band), cypermethrin (Cyp; 3–65 µg/band), phorate (Pho; 1–10 µg/band), (b) cyprodinil (Cypr; 3–25 µg/band), mercaptodimethur (Mer; 1–20 µg/band), and chlorpyrifos (Chl; 3–25 µg/band), separated on HPTLC plate silica gel 60 RP-18 W with n-hexane/ethyl acetate 5:1 (V/V) up to 60 mm

Fig. 2

Dose–response curves (Table S2) exemplarily determined for the estrogenic responses of seven pesticides obtained from biodensitograms on three different days (n = 3; standard deviation as error bar)

Fig. 3

PPP screening: HPTLC–pYES–FLD bioautogram image at 366 nm of six commercial PPPs (Table S3) showing blue fluorescent estrogenic responses but also opposing effects (dark bands marked with asterisk indicate true anti-estrogens or false-positive anti-estrogens or cyctotoxins) on HPTLC plate silica gel 60 RP-18 W, developed with n-hexane/toluene/ethyl acetate 4:1:1 (V/V/V) up to 70 mm. Dit, Dithane^® NeoTec (mancozeb, 2.3 µg/band); Dyn, DYNALI^® (difenoconazol and cyflufenamid, 0.2 and 0.9 µg/band, respectively); Fol, Folpan^® 80 WDG (folpet, 2.4 µg/band); Swi, SWITCH^® (fludioxonil and cyprodinil, 0.8 and 1.1 µg/band, respectively); Tel, Teldor^® (fenhexamid, 1.5 µg/band); Viv, Vivando^® (metrafenone, 1.5 µg/band)

Fig. 4

White wine screening: (a) HPTLC–FLD chromatogram at 366 nm and (b–d) corresponding HPTLC–pYES–FLD bioautograms of the Rivaner white wine sample (W, 35 µL/band) showing the blue fluorescent estrogenic zones 1–4 (only slight opposing effects evident) in comparison to the pesticides (S) fludioxonil (Flu; 1.2 µg/band), fenhexamid (Fen; 1 µg/band), and cyprodinil (Cyp; 30 µg/band) as (b) separate tracks or (c, d) overlapped with the sample, which revealed (c) fenhexamid residues in zone 1 and possibly also (d) cyprodinil, separated on HPTLC plate silica gel 60 RP-18 W with n-hexane/ethyl acetate 5:1 (V/V) up to 60 mm

Fig. 5

White seedless table grape skin screening: (a) HPTLC–FLD chromatogram at 366 nm and (b-d) corresponding HPTLC–pYES–FLD bioautograms of the skin of a grape from Chile (Ch, 40 µL/band) showing blue fluorescent estrogenic zones 5 and 6, which revealed fenhexamid residues in zone 5, whereas zone 6 was present in all 18 grape samples studied, in comparison to the pesticides (S) fludioxonil (Flu; 1.2 µg/band), fenhexamid (Fen; 1 µg/band), and cyprodinil (Cyp; 30 µg/band) as (b) separate track or (c, d) overlapped with the sample, separated on HPTLC plate silica gel 60 RP-18 W with n-hexane/toluene/ethyl acetate 5:1:1 (V/V/V) up to 70 mm. (e) Further screening (with higher elution power using n-hexane/toluene/ethyl acetate 4:1:1, V/V/V) of samples from Brazil (B, Festival Seedless bought from Penny_P and Sugar Crispy from REWE_R), Peru (P, Prime Seedless from Lidl), and Italy (I, from local market), detected at white light illumination and FLD 366 nm, all 60 µL/band

Fig. 6

Tomato screening: (a) HPTLC–FLD chromatogram at 366 nm and (b) corresponding HPTLC–pYES–FLD bioautogram of pulp (tp) and skin (ts) of a cherry tomato from Belgium (40 µL/band, bought from Penny) showing the blue fluorescent estrogenic zone 7 present in all samples, in comparison to the pesticides (S) fludioxonil (Flu; 1.2 µg/band), fenhexamid (Fen; 1 µg/band), and cyprodinil (Cyp; 30 µg/band) separated as in Fig. 5

Fig. 7

Household tomato washing experiment: (a) HPTLC–pYES–FLD bioautograms at 366 nm of tomato pulp (Tp, 40 µL/band) or (b) tomato skin (Ts, 40 µL/band, same tomato sample as in Fig. 6) treated either with fludioxonil (Flu; 4 µg/halved tomato), fenhexamid (Fen; 4 µg/halved tomato), or cyprodinil (Cyp; 20 µg/halved tomato) (a, b) before extraction, or (c) spiked (_s) post-extraction and (b) skin wiped with a cotton swab (_W) in comparison to negative control (NC) and pesticide standards, separated as in Fig. 5. Zone 7 increased in the response by the treatment with cyprodinil (7*)