- Email: [email protected]
Separation and identification of hormone-active compounds using a combination of chromatographic separation and yeast-based reporter assay
Science of the Total Environment 605–606 (2017) 507–513
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Separation and identification of hormone-active compounds using a combination of chromatographic separation and yeast-based reporter assay Alexandre Chamas a, Ha Thi Minh Pham a,b, Martin Jähne c, Karina Hettwer c, Linda Gehrmann d, Jochen Tuerk d, Steffen Uhlig c, Kirsten Simon e, Kim Baronian f, Gotthard Kunze a,⁎ a
Leibniz Institute of Plant Genetics and Crop Plant Research (IPK), Corrensstr. 3, D-06466 Gatersleben, Germany Institute of Biotechnology, Vietnam Academy of Science and Technology, 18 Hoang Quoc Viet, Cau Giay, Ha Noi, Viet Nam c QuoData GmbH, Prellerstr. 14, D-01309 Dresden, Germany d Institut für Energie- und Umwelttechnik e. V. (IUTA, Institute of Energy and Environmental Technology), Bliersheimer Str. 58-60, D-47229 Duisburg, Germany e New diagnostics GmbH, Moosstr. 92c, D-85356 Freising, Germany f School of Biological Sciences, University of Canterbury, Private Bag 4800, Christchurch, New Zealand b
H I G H L I G H T S
G R A P H I C A L
A B S T R A C T
• A combination of thin-layer chromatography and yeast-based assay is presented. • This assay can detect picogram amounts of estrogens, androgens and progestogens. • Environmental samples were successfully assayed.
a r t i c l e
i n f o
Article history: Received 21 April 2017 Received in revised form 9 June 2017 Accepted 9 June 2017 Available online xxxx Editor: D. Barcelo Keywords: Steroid hormones Thin-layer chromatography Yeast assay Fluorescence Wastewater samples
a b s t r a c t Arxula adeninivorans-based yeast cell assays for the detection of steroid hormones demonstrated their efficiency for the determination of total hormone activity in a variety of samples using a microtiter plate format. In this study, a preliminary chromatographic separation using thin-layer chromatography plates is introduced in order to allow a rapid identification of the compounds responsible for this hormonal activity. The yeast whole cell assay can then be performed on the plate, producing a detectable signal where a steroid hormone is present. Simultaneous detection of estrogens, progestogens and androgens on the same plate in the picogram range was achieved, while keeping the assay as simple and affordable as possible. The assay requires a single incubation of the thin-layer chromatography plate and the detection of reporter protein production can be performed by fluorescence scanning of the plate at different wavelengths. The chromatographic separation allows the separation of several estrogens, androgens and progestogens, thus making its application for ‘real world’ samples very useful. In this work, different water-based samples from environmental origins were used to demonstrate the capacity of this new bioassay. Trials showed that most samples, with the exception of complex samples such as wastewater influent, can be assayed. © 2016 Elsevier B.V. All rights reserved.
⁎ Corresponding author at: Leibniz Institute of Plant Genetics and Crop Plant Research (IPK), Corrensstr. 3, D-06466 Gatersleben, Saxony-Anhalt, Germany. E-mail address: [email protected] (G. Kunze).
http://dx.doi.org/10.1016/j.scitotenv.2017.06.077 0048-9697/© 2016 Elsevier B.V. All rights reserved.
508
A. Chamas et al. / Science of the Total Environment 605–606 (2017) 507–513
1. Introduction Yeast-based bioassays, which can detect the biological activity of many compounds in the environment, play an important role in the field of bioanalytics. A significant proportion of these bioassays are dedicated to the detection of steroid hormones (Adeniran et al., 2015) as these can identify endocrine disrupting compounds (EDCs), compounds which can bind to hormone receptors and disturb the vertebrate endocrine system (Kabir et al., 2015; Casals-Casas and Desvergnes, 2011). There are yeast-based bioassays designed to detect estrogens (Kaiser et al., 2010; Sanseverino et al., 2008), progestogens (Viswanath et al., 2008; Chamas et al., 2015), glucocorticoids (Bovee et al., 2011; Pham et al., 2016), androgens (Bhattacharjee and Khurana, 2014; Gerlach et al., 2014) and more recently, for the simultaneous detection of several hormone classes (Chamas et al., 2017). Most of these bioassays combine the production of recombinant human hormone receptors and use enzymes or fluorescent reporter proteins to produce a quantitative signal. Contrary to standard analytical detection methods which give the exact composition of a sample, yeast-based bioassays quantify the total hormonal activity of a sample and therefore assess the biological impact a sample can have. Yeast-based bioassays are thus often used as a pre-screening procedure to give an overview of a samples hormonal activity before analytical detection methods are used to determine which compounds are present in the sample and identify potential EDCs. In an effort to combine the bioassay and analytical approaches, recent studies have introduced a bioautography detection method using thin-layer chromatography and cells from the yeast-based estrogen screen (YES) assay (Mueller et al., 2004; Spira et al., 2013). In these tests, an unknown sample is firstly separated on a thin-layer chromatography plate together with control molecules with known estrogenic activity. After separation, the plate is sprayed with an engineered yeast cell suspension producing a β-galactosidase when an estrogenic substance binds to the recombinant human estrogen receptor. The presence of this enzyme is then detected by spraying the plate again with a chromogenic or fluorescence-producing substrate. With this planarYES assay, compounds with estrogenic activity have been detected in various environmental samples (Klingelhöfer and Morlock, 2015; Buchinger et al., 2013), cosmetic products (Buchinger et al., 2013) and food samples (Klingelhöfer and Morlock, 2014). Detection of 17βestradiol in the femtogram range was reported in environment samples, thus enabling ultratrace level determination. A great advantage of TLC based assays over regular microtiter plate based assays is that prior separation of a complex sample during thin-layer chromatography will remove compounds that could be toxic to yeast. Another advantage of the planar-YES assay is that it requires less than 24 hours incubation to obtain a signal after chromatographic separation whereas the regular YES assay requires incubation for 3 to 5 days to obtain an unequivocal result (Routledge and Sumpter, 1996). To obtain sharp bands and very low detection limits, the latest optimizations of the planar-YES assay employs high performance thin layer chromatography plate coatings such as Reverse phase RP-18W (Klingelhöfer and Morlock, 2014) and fluorescence-producing substrate 4-methylumbelliferyl-b-D-galactopyranoside (MUG) (Buchinger et al., 2013). These improvements make the planar-YES assay of comparable performance to the established analytical methods, offering determination of a samples biological activity at low concentrations as well as identification of the compounds. However, the assay is more complex and more expensive than previous assays which may reduce the attractiveness of this analytical/biological detection method. In an effort to incease simplicity in preference to low sensitivity and a broad detection range, we designed a bioassay combining analytical separation with unmodified silica gel thin-layer chromatography plates and detection without addition of an enzyme substrate. To achieve the latter goal, fluorescent reporter proteins were used, obviating the need for substrate. Moreover, ideal EDC detection should not only detect estrogenic molecules but also other EDCs such as progestogens
and androgens. Therefore our aim was to design a bioassay able to determine in one analysis which endocrine activity is present in a sample and which compounds are responsible for it. This assay, to our knowledge, is the first to possess these characteristics and could be developed as an assay for other molecules. Modified Arxula adeninivorans strains were used as the biocomponent in the assay. These strains, described in a recently published research article (Chamas et al., 2017), produce a recombinant human progesterone, androgen or estrogen receptor, each in combination with the production of a fluorescent reporter protein, CFP, GFP or DsRed2 respectively. In the microtiter plate assay, these strains were mixed and could then, once incubated with a mixture of estrogens, androgens and progestogens, determine which particular hormone activity is present in the sample. It was then decided to test if this yeast strain mix can be combined with thin-layer chromatography and after three fluorescence scans of a plate, determine which hormone activities are present in a sample. To achieve this, the bioassay should realise these objectives: (1) all three A. adeninivorans strains should be able to interact with their target hormone on a thin-layer chromatography plate, (2) the fluorescence emitted from this interaction should be detectable with a plate scanner, (3) the three yeast strains, when mixed, should give easily distinguishable signals for each of the three hormonal types, (4) the development conditions for the assay should be uncomplicated. 2. Material and methods 2.1. Chemicals Estrone (E1), 17β-estradiol (E2), estriol (E3), progesterone (P), medroxyprogesterone acetate (MP), mifepristone (M), nandrolone (N), trenbolone (T), 5α-dihydrotestosterone (DHT) were all purchased from Sigma-Aldrich (Steinheim, Germany). All hormones were solubilized in ethanol at a stock concentration of 1 g/l. Acetone, dichloromethane and cyclohexane were all also purchased from Sigma-Aldrich (Steinheim, Germany). 2.2. Strains and cultivation conditions The yeast strains used in this work, (A. adeninivorans G1212/ YRC102-hPR-CFP, G1212/YRC102-hAR-GFP and G1212/YRC102-hERDsRed2), are described in a previous publication (Chamas et al., 2017). Each strain was cultivated for 24 h in yeast minimal media supplemented with glucose (Yeast Minimal Medium [YMM]-glucose) at 30 °C until the OD620 nm reached approximately 3. The cells were pelleted by centrifugation at 4000 × g for 10 min, the pellet was resuspended in an equal volume of fresh YMM-maltose and the three cell suspensions were mixed together to create A. adeninivorans G1212/YRC102-hHRfluo at a final total OD620 nm of 3. 2.3. Thin-layer chromatography procedure 100 × 200 mm polyester sheets precoated with 0.2 mm silica gel and without fluorescent markers (Macherey-Nagel, Düren, Germany) were used for all thin-layer chromatography experiments. For automated application of samples, the desired volumes were applied using an automatic TLC Sampler 4 (CAMAG, Muttenz, Switzerland) as 6 mm bands, 8 mm from the bottom of the plate. The application zones were allowed to dry at room temperature for 5 min and the plate was inserted in a TLC vertical development chamber containing dichloromethane:cyclohexane:acetone (60:35:5, v/v/v). Once the solvent migrated to within 85 mm from the bottom of the plate, the plate was removed from the developing chamber and allowed to dry at room temperature for 30 min. For manual sample application, the desired volume was applied with a glass capillary at 10 mm from the bottom of the plate. After drying for 5 min at room temperature, the plate was focused two times with methanol until the solvent
A. Chamas et al. / Science of the Total Environment 605–606 (2017) 507–513
reached 20 mm so that sharp bands formed before development with the solvent. Between each focusing, the plate was dried at 50 °C for 5 min to completely remove methanol. Vertical development occurred until the dichloromethane:cyclohexane:acetone solvent reached 85 mm from the bottom of the plate.
2.4. Chemical detection Developed plates were sprayed with 8% sulfuric acid in ethanol and incubated for 15 min at 110 °C. Absorbance at 310 nm and 500 nm of each track was then measured with a TLC Scanner 3 (CAMAG, Muttenz, Switzerland). The CAMAG-embedded software, Wincats was used to provide a graphical display of the absorbing zones and calculate the peak areas of the peaks.
2.5. Bioassay detection The developed plates were immersed in A. adeninivorans G1212/ YRC102-hHR-fluo for 5 s and the excess cell suspension was removed with a clean paper towel. The plate was then inserted in a closed plastic box with wet towel and the box was incubated for 18 h at 30 °C. The plate was dried at room temperature for 30 min and three fluorescence scans of each track were performed to determine CFP, GFP and DsRed2 fluorescence. The excitation wavelengths and emission filters were respectively 445 nm/K460 nm for CFP, 485 nm/K500 nm for GFP and 542 nm/K560 nm for DsRed2.
509
2.6. Wastewater samples The bioassay was trialed with environmental samples from seven different origins: stream (Saxony, Germany), source water (Brandenburg, Germany and Saxony, Germany), river (Saxony, Germany), lake (Berlin, Germany), tap water (Berlin, Germany), wastewater plant influent (North Rhine-Westphalia, Germany) and wastewater plant effluent (Rhineland-Palatinate, Germany). Precise locations of these sources can be provided upon request. Samples were not pretreated or modified prior to the assay. To assess the suitability of the combined TLC assay for environmental analysis, a mix of the three hormones E2, P and DHT (all at the concentration 1 μg/l) was added to double distilled water (positive control) and to all samples. Column-mediated extraction of the positive control and the samples was performed with an enrichment factor of 20. Each sample (5 μl volume) was then manually applied to the TLC plates for bioassay detection. After the three fluorescence scans, the area of the different peaks was calculated with Wincats to allow comparison between the positive control and the samples. 3. Results and discussion 3.1. Determination of minimum detectable amount of each hormone class Different quantities of the three hormones progesterone (P), 5αdihydrotestosterone (DHT) and 17β-estradiol (E2) solubilized in 5 μl ethanol, were applied separately to three thin-layer chromatography plates and submitted to biological detection with the three strains of A. adeninivorans without chromatographic separation. At the end of
Fig. 1. (A) Densitogram of a TLC plate incubated with A. adeninivorans G1212/YRC102-hER-DsRed2 where different amounts of E2 were applied. The fluorescence scan was performed at 542 nm (excitation) with a K560 nm filter. The relationship between peak area and amount of hormone applied to TLC plates and incubated with A. adeninivorans G1212/YRC102-hERDsRed2, G1212/YRC102-hAR-GFP and G1212/YRC102-hPR-CFP is shown in (B), (C) and (D), respectively.
510
A. Chamas et al. / Science of the Total Environment 605–606 (2017) 507–513
the 18 hours incubation, the three plates were scanned at the appropriate wavelength for the reporter protein concerned to obtain three different densitograms. An example of densitogram obtained when a TLC plate with different quantities of E2 was incubated with A. adeninivorans G1212/YRC102-hER-DsRed2 is visible in Fig. 1A. A clear peak is visible for amounts of hormone superior to 7.5 pg. After determining the area of the peak via the CAMAG software, Wincats, a diagram illustrating the relationship between the amount of applied E2 and the peak area was created and can be seen on Fig. 1B. Additionally, the same diagrams were created for the plate incubated with A. adeninivorans G1212/YRC102-hPR-CFP (Fig. 1C) and G1212/ YRC102-hAR-GFP (Fig. 1D). In all three cases, a clear increase in the peak area with increasing hormone amount following a 4-parameters logarithmic curve can be seen. By reproducing these experiments, it appeared that even though the minimal detectable amount of each hormone remained the same, significant differences between the peak areas for identical quantities of hormone were seen on different plates (data not shown). Thus, in all cases, similar 4-parameters logarithmic curves were seen but with different maximum values. To explain this phenomenon, one hypothesis is that thin-layer chromatography plates vary from batch to batch in terms of silica coating thickness and quality. As it was already shown that yeast actually grow on silica plate (Klingelhöfer and Morlock, 2014), these structural differences of the plate can lead to yeast growth variations and therefore to higher or lower fluorescence. From all experiments, minimal detectable amounts of 7.5 pg E2, 25 pg P and 25 pg DHT were detected, which correspond to initial concentrations applied to the plates of 1.5 μg/l, 5 μg/l and 5 μg/l, respectively. Even though these concentrations seem high when compared to the detection limits of microtiter plate based yeast assays, it is important to understand that more than 5 μl of an ethanol-based hormone dilution can be applied in the same track, thus allowing the utilization of lower dilutions than 1.5 μg/l E2, 5 μg/l progesterone and 5 μg/l DHT which were made for each hormone. Additionally, because ethanol evaporates quickly, successive applications of this hormone can be performed with intermediary drying phase, thus lowering again the bioassay detection limit. Another important point which can extracted from
Fig. 1A is that, even though the peaks get slightly broader with increasing E2 amount, this effect remains small when compared to the results obtained with a similar plate in the paper of Mueller et al. (2004). This indicates that the phenomenon responsible for peak diffusion in silicabased TLC plates without chromatographic separation may not be the long incubation with water-based cell suspension but rather the second incubation with the water-based substrate solution at higher temperature (37 °C). 3.2. Separation of estrogens, androgens and progestogens by TLC – chemical detection Thin-layer chromatography, due to the relatively small plates, allows the separation of only a limited number of compounds and this separation occurs on the basis of the compounds adsorption properties on the solid phase as well as its solubility in the mobile phase. These physical properties are frequently related to the atomic structure of the molecule and consequently, similar molecules will be harder to separate in a TLC format. Steroid hormones, all originating from the molecule cholesterol, have a similar structure and therefore several solvents were trialed to find an acceptable universal mobile phase for estrogens, progestogens and androgens. The best separation was obtained with a mix of dichloromethane, cyclohexane and acetone with the volume proportions 60:35:5. To show this separation, three estrogens (estrone (E1), E2 and estriol (E3)), three progestogens (P, medroxyprogesterone acetate (MP) and mifepristone (M)) and three androgens (DHT, nandrolone (N) and trenbolone (T)) as well as mixtures of these estrogens, progestogens and androgens were applied to a TLC plate; 360 ng of each compound per track. After separation, an non-specific chemical detection was performed by spraying 8% sulfuric acid and incubating the plate at 110 °C for 15 min. Strong acid is known to oxidize a very wide range of molecules and steroid hormones in particular. This oxidation produces a dark stain on the TLC plate (Bhawani et al., 2010) which can be scanned at 500 nm, with the exception of oxidized progesterone which is only visible under UV. The densitogram obtained after TLC plate scan is visible in Fig. 2. Each visible peak was then attributed to a
Fig. 2. Densitogram of a TLC plate where several hormones of different classes were applied and allowed to migrate in dichloromethane:cyclohexane:acetone solvent. Track 1: E1, track 2: E2, track 3: E3, track 4: mix of E1, E2 and E3, track 5: P, track 6: MP, track 7: M, track 8: mix of P, MP and M, track 9: DHT, track 10: N, track 11: T, track 12: mix of DHT, N and T. The scan was performed at 500 nm after chemical detection.
A. Chamas et al. / Science of the Total Environment 605–606 (2017) 507–513 Table 1 Retention factor of selected hormones determined with chemical detection. Hormone class
Hormone
Rf
Estrogens
Estrone 17β-Estradiol Estriol Progesterone Medroxyprogesterone acetate Mifepristone 5α-Dihydrotestosterone Nandrolone Trenbolone
0.49 0.26 0.03 0.53 0.51 0.28 0.37 0.24 0.01/0.23
Progestogens
Androgens
specific hormone and the retention factor of each compound was determined. These values can be found in Table 1. For progesterone, the plate was scanned at 310 nm to determine the retention factor (TLC scan not shown). A good separation of the three estrogens E1, E2 and E3 is visible with this mobile phase with retention factors separated by at least 0.2. P and MP also show similar retention factor values as would be expected from
511
the similarity of these compounds however the third progestogen, mifepristone, has a much lower retention factor. The three androgens show relatively good separation with the particularity that T presents two peaks, one with a retention factor of 0.01 and the second by 0.23, probably indicating that the commercial T consists in two isoforms with two different chromatographic properties. It is also important to note that the three most common steroid hormones - E2, P and DHT – show well separated retention factors. 3.3. Separation of estrogens, androgens and progestogens by TLC – fluorescent detection It has been shown that the three strains of A. adeninivorans G1212/ YRC102-hPR-CFP, G1212/YRC102-hAR-GFP and G1212/YRC102-hERDsRed2 can be simultaneously incubated with hormones from the three different classes in microtiter plate format (Chamas et al., 2017) and give three distinct signals. To test whether such incubation can also be applied to the combination assay, 100 pg of E2, P and DHT and a mix of these three hormones was applied to a TLC plate and separated with dichloromethane:cyclohexane:acetone mobile phase. After
Fig. 3. Densitogram of a TLC plate allowed to migrate in dichloromethane:cyclohexane:acetone solvent where E2 (violet line), P (blue line) and DHT (green line) as well as a mix of these hormones (yellow line) were applied. Fluorescence scans at the wavelength 445 nm/K460 nm for CFP (A), 485 nm/K500 nm for GFP (B) and 542 nm/K560 nm for DsRed2 (C) were performed after incubation of the plate with A. adeninivorans G1212/YRC102-hHR-fluo.
512
A. Chamas et al. / Science of the Total Environment 605–606 (2017) 507–513
development, the plate was dipped into a suspension of A. adeninivorans G1212/YRC102-hPR-CFP, G1212/YRC102-hAR-GFP and G1212/ YRC102-hER-DsRed2. After 18 hours incubation the plate was scanned at three different wavelengths to detect CFP, GFP and DsRed2 fluorescence as described in Material and Methods. The densitograms obtained for these three scans can be seen in Fig. 3. For each densitogram, the fluorescence of the four tracks were recorded and plotted on the same diagram with blue, green, violet and yellow curves corresponding respectively to the tracks where P, DHT, E2 and a mixture of the three were applied. In comparison to the detection without chromatographic development, the shape of these peaks is more disturbed and the baseline is not completely flat. Nevertheless, peaks are present in all three scans and by comparing the Rfs of these peaks with the Rfs determined in the preceding section, all three hormones can be successfully detected. When the plate was scanned for GFP fluorescence (Fig. 3.B), the peak of E2 is clearly visible in the track where E2 only was applied and in the track with a mix of the three hormones. This could indicate that the A. adeninivorans G1212/YRC102-hAR-GFP cells can detect E2, but previous work showed that the androgen receptors affinity for E2 is very low (Gerlach et al., 2014; Scippo et al., 2002). To confirm this, an additional strain producing the androgen receptor and possessing CFP as reporter protein was created and incubated with a TLC plate where E2 was applied. This experiment showed no E2 peak at all (data not shown). To explain the GFP fluorescence observed at the Rf of E2, the most plausible answer is found in the fluorescence detector. The detector of the CAMAG scanner 3 is preceded by fluorescence filters which will cut all light of wavelengths shorter than the desired wavelength, thus only allowing detection of the emitted light and not the excitation light. In the case of the GFP fluorescence detection, in order to excite A. adeninivorans G1212/YRC102-hAR-GFP cells at 485 nm and measure fluorescence at 520 nm, the K500 filter was used to remove emission light with wavelengths below 500 nm before it strikes the detector. It was shown by Bevis and Glick (2002) that DsRed possesses two excitation peaks, one small at 488 nm and a larger one at 558 nm, and one emission peak at 583 nm. Therefore, the 485 nm excitation will also induce dsRED fluorescence through the small excitation peak and the emitted light from this fluorescent protein will not be removed by the K500 filter. To overcome this phenomenon, two main strategies can
be followed. Firstly the DNA sequence of the DsRed2 protein can be genetically modified to remove the excitation peak at 488 nm. Secondly, the utilization of band filters in the Scanner detector could allow more precise selection of which emitted wavelength can be observed. The use of such a filter is not possible with the CAMAG Scanner 3 but other models offer this possibility. 3.4. Application of wastewater samples – fluorescent detection One of the applications of the combined bioassay is to detect hormone activity in environmental samples. However it has already been shown that such samples generally have relatively low concentrations of steroid hormones, mostly in the ng/l range for estrogens for example. Consequently, it was decided to use spiked samples as described in the Materials and methods section. The same mix of E2, P and DHT was added to either the environmental samples or double distilled water (the positive control) and, after extraction, 5 μl of the eluted volume was applied to a TLC plate. After the three scans for CFP, GFP or DsRed2 fluorescence, the peak areas for each hormone in each sample and in the positive control were determined. To compare the hormone activity of each sample, a simple normalization was performed by dividing the peak area of one hormone in a sample by the peak area of the same hormone in the positive control. A value of 1 indicates that the same hormone concentration was detected in the sample and in the positive control. This allows assessment of whether or not the sample matrix has an effect on the detection of hormones. The same test was performed on two different TLC plates on different days and the results of these two experiments were used for statistical analysis. The relative hormone activity of each sample for the three hormones is presented in the three diagrams in Fig. 4 with error bars representing the 95% confidency interval of the measurement. The samples, source 1, river 1, river 2, tap water and wastewater effluent show values around 1 for each three hormones, thus indicating that the spiked concentration of hormone was producing similar peaks in the positive control and in the samples. The sample stream presents ratios of 1 for E2 and DHT but for P, the ratio is slightly higher. In the case of samples source 2 and wastewater influent 2, all three hormones have been detected but all ratio are inferior to 1, indicating a smaller area of the hormone
Fig. 4. Ratio values of androgens (blue bars), estrogens (red bars) and progestogens (green bars) in different environmental samples. Ratio value is calculated by dividing the peak area of a given hormone in a spiked sample by the peak area of the same hormone in the positive control. Error bars indicate the 95% confidence interval of the ratio.
A. Chamas et al. / Science of the Total Environment 605–606 (2017) 507–513
peak in those samples if compared to the positive control. Finally, the lake and wastewater influent 1 samples show high androgens and estrogen ratio but with large error bars. This wide 95% confidence interval shows that, for these two types of samples, the reliability of the measurement is low, thus indicating that the detection of endocrine active compounds in such complex samples with the described assay has to be improved. The reason for this low reliability may be that additional compounds which are not hormones but which are extracted simultaneously, interfere with the yeast metabolism or the fluorescence molecule. The same limitations were also seen with a yeast assay in microtiter plate format where environmental samples with complex matrices were found to be cytotoxic in many cases (Gehrmann et al., 2016). The fact that detecting estrogens in wastewater treatment plants influents is possible has been demonstrated by Klingelhöfer and Morlock (2015) and therefore the limitations observed in the present study regarding this type of sample may have to be overcome with an improved extraction process. Taken together, these results indicate that for samples with relatively simple composition and good purity like tap water or wastewater effluent, the separation and determination of the three steroid hormones is possible. However, for more complicated samples such as wastewater treatment plant influent, the bioassay protocol has to be optimized. 4. Conclusions The bioassay developed here enables the separation and a rapid identification of hormones of different classes within one day. The use of A. adeninivorans G1212/YRC102-hHR-fluo biocomponent allows simultaneous detection of estrogens, androgens and progestogens on one plate with only a fluorescence plate reader to perform the assay. Because the potential applications of this combined assay are in environmental science, a field where fully-equipped laboratories with analytical measurement devices are not always available, special attention was given to keep the assay as simple as possible. Simple TLC plates with silica matrix, common solvents and manual sample application are all that is required for the standard procedure which allowed the detection of well-separated peaks. The use of fluorescence reporter proteins removes the need of a second incubation step, which is necessary in the case of enzyme-mediated detection, and allows the detection of the signal in less than 24 h. However the assays limit of detection needs to be reduced in order to identify and quantify EDCs present in environmental samples. Future work will therefore focus on identifying or developing a simple extraction process for complex samples and development of yeast strains with improved fluorescence reporters. With this bioassay, EDCs can also be identified by simply comparing retention factors of active compounds with the retention factors of known hormones. In future work, establishment of a method to remove active compounds by scratching from the thin-layer chromatography plate for subsequent analytic detection by HPLC could provide a method to corroborate the identity of molecules identified by the assay and also potentially allow the discovery of new EDCs. Acknowledgements The work was supported by grant from the German Federal Ministry of Economics and Technology (Bundesministerium für Wirtschaft und Technologie, BMWi) within its program “Zentrales Innovationsprogramm
513
Mittelstand, ZIM” (Grant Numbers KF2131626SB4, KF2715706SB4, KF2025438SB4). References Adeniran, A., Sherer, M., Tyo, K.E.J., 2015. Yeast-based biosensors: design and applications. FEMS Yeast Res. 15, 1–15. Bevis, B.J., Glick, B.S., 2002. Rapidly maturing variants of the Discosoma red fluorescent protein (DsRed). Nat. Biotechnol. 20, 83–87. Bhattacharjee, G.B., Khurana, S.M.P., 2014. In vitro reporter assays for screening of chemicals that disrupt androgen signalling. J. Toxicol. 2014, ID 701752. http://dx. doi.org/10.1155/2014/701752. Bhawani, S.A., Sulaiman, O., Hashim, R., Mohamad Ibrahim, M.N., 2010. Thin-layer chromatographic analysis of steroids: a review. Trop. J. Pharm. Res. 9, 301–313. Bovee, T.F., Helsdingen, R.J., Hamers, A.R., Brouwer, B.A., Nielen, M.W., 2011. Recombinant cell assay for the detection of (gluco)corticosteroids and endocrine-disrupting potencies of several environmental PCB contaminants. Anal. Bioanal. Chem. 401, 873–882. Buchinger, S., Spira, D., Bröder, K., Schlüsener, M., Ternes, T., Reifferscheid, G., 2013. Direct coupling of thin-layer chromatography with a bioassay for the detection of estrogenic compounds, application for effect-directed analysis. Anal. Chem. 85, 7248–7256. Casals-Casas, C., Desvergnes, B., 2011. Endocrine disruptors: from endocrine to metabolic disruption. Annu. Rev. Physiol. 73, 135–162. Chamas, A., Nieter, A., Pham, H.T.M., Giersberg, M., Hettwer, K., Uhlig, S., Simon, K., Baronian, K., Kunze, G., 2015. Development of a recombinant Arxula adeninivorans cell bioassay for the detection of molecules with progesterone activity in wastewater. Anal. Bioanal. Chem. 407, 8109–8120. Chamas, A., Pham, H.T.M., Jähne, M., Hettwer, K., Uhlig, S., Simon, K., Einspanier, A., Baronian, K., Kunze, G., 2017. Simultaneous detection of three sex steroid hormone classes using a novel yeast-based biosensor. Biotechnol. Bioeng. 114 (2017), 1539–1549. Gehrmann, L., Bielak, H., Behr, M., Itzel, F., Lyko, S., Simon, A., Kunze, G., Dopp, E., Wagner, M., Tuerk, J., 2016. (Anti-)estrogenic and (anti-)androgenic effects in wastewater during advanced treatment: comparison of three in vitro bioassays. Environ. Sci. Pollut. Res. Int. 1–11. Gerlach, T.K., Kaiser, C., Körner, M., Hettwer, K., Uhlig, S., Simon, K., Baronian, K., Kunze, G., 2014. Development and assessment of a novel Arxula adeninivorans androgen screen (A-YAS) assay and its application in analysis of cattle urine. Sci. Total Environ. 490, 1073–1081. Kabir, E.R., Rahman, M.S., Rahman, I., 2015. A review on endocrine disruptors and their possible impacts on human health. Environ. Toxicol. Pharmacol. 40, 241–258. Kaiser, C., Uhlig, S., Gerlach, T., Körner, M., Simon, K., Kunath, K., Florschütz, K., Baronian, K., Kunze, G., 2010. Evaluation and validation of a novel Arxula adeninivorans estrogen screen (nAES) assay and its application in analysis of wastewater, seawater, brackish water and urine. Sci. Total Environ. 408, 6017–6026. Klingelhöfer, I., Morlock, G.E., 2014. Sharp-bounded zones link to the effect in planar chromatography-bioassay-mass spectrometry. J. Chromatogr. A 1360, 288–295. Klingelhöfer, I., Morlock, G.E., 2015. Bioprofiling of surface/wastewater and bioquantitation of discovered endocrine-active compounds by streamlined direct bioautography. Anal. Chem. 87, 11098–11104. Mueller, M.B., Dausend, C., Weins, C., Frimmel, F.H., 2004. A new bioautographic screening method for the detection of estrogenic compounds. Chromatographia 60, 207–211. Pham, H.T.M., Chamas, A., Nieter, A., Giersberg, M., Rutten, T., Gehrmann, L., Hettwer, K., Tuerk, J., Uhlig, S., Simon, K., Baronian, K., Kunze, G., 2016. Determination of glucocorticoids using photometric (A-YGS) and spectrofluorometric (A-YGFS) bioassays based on modified Arxula adeninivorans cells: applications in environmental analysis. Sensors Actuators B Chem. 223, 540–549. Routledge, E.J., Sumpter, J.P., 1996. Estrogenic activity of surfactants and some of their degradation products assessed using a recombinant yeast screen. Environ. Toxicol. Chem. 15, 241–248. Sanseverino, J., Elridge, M.L., Layton, A.C., Easter, J.P., Yarbrough, J., Schultz, T.W., Sayler, G.S., 2008. Screening of potentially hormonally active chemicals using bioluminescent yeast bioreporters. Toxicol. Sci. 107, 122–134. Scippo, M.L., Van de Weerdt, C., Willemsen, P., Francois, J.M., Rentier-Delrue, F., Müller, M., Martial, J.A., Maghuin-Rogister, G., 2002. Detection of illegal growth promoters in biological samples using receptor binding assays. Anal. Chim. Acta 473, 135–141. Spira, D., Reifferscheid, G., Buchinger, S., 2013. Combination of high-performance thinlayer chromatography with a specific bioassay – a tool for effect-directed analysis. J. Planar Chromatogr.–Mod. TLC 26, 395–401. Viswanath, G., Halder, S., Divya, G., Majumder, C.B., Roy, P., 2008. Detection of potential (anti)progestagenic endocrine disruptors using a recombinant human progesterone receptor binding and transactivation assay. Cell. Mol. Endocrinol. 295, 1–9.
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Separation and identification of hormone-active compounds using a combination of chromatographic separation and yeast-based reporter assay Alexandre Chamas a, Ha Thi Minh Pham a,b, Martin Jähne c, Karina Hettwer c, Linda Gehrmann d, Jochen Tuerk d, Steffen Uhlig c, Kirsten Simon e, Kim Baronian f, Gotthard Kunze a,⁎ a
Leibniz Institute of Plant Genetics and Crop Plant Research (IPK), Corrensstr. 3, D-06466 Gatersleben, Germany Institute of Biotechnology, Vietnam Academy of Science and Technology, 18 Hoang Quoc Viet, Cau Giay, Ha Noi, Viet Nam c QuoData GmbH, Prellerstr. 14, D-01309 Dresden, Germany d Institut für Energie- und Umwelttechnik e. V. (IUTA, Institute of Energy and Environmental Technology), Bliersheimer Str. 58-60, D-47229 Duisburg, Germany e New diagnostics GmbH, Moosstr. 92c, D-85356 Freising, Germany f School of Biological Sciences, University of Canterbury, Private Bag 4800, Christchurch, New Zealand b
H I G H L I G H T S
G R A P H I C A L
A B S T R A C T
• A combination of thin-layer chromatography and yeast-based assay is presented. • This assay can detect picogram amounts of estrogens, androgens and progestogens. • Environmental samples were successfully assayed.
a r t i c l e
i n f o
Article history: Received 21 April 2017 Received in revised form 9 June 2017 Accepted 9 June 2017 Available online xxxx Editor: D. Barcelo Keywords: Steroid hormones Thin-layer chromatography Yeast assay Fluorescence Wastewater samples
a b s t r a c t Arxula adeninivorans-based yeast cell assays for the detection of steroid hormones demonstrated their efficiency for the determination of total hormone activity in a variety of samples using a microtiter plate format. In this study, a preliminary chromatographic separation using thin-layer chromatography plates is introduced in order to allow a rapid identification of the compounds responsible for this hormonal activity. The yeast whole cell assay can then be performed on the plate, producing a detectable signal where a steroid hormone is present. Simultaneous detection of estrogens, progestogens and androgens on the same plate in the picogram range was achieved, while keeping the assay as simple and affordable as possible. The assay requires a single incubation of the thin-layer chromatography plate and the detection of reporter protein production can be performed by fluorescence scanning of the plate at different wavelengths. The chromatographic separation allows the separation of several estrogens, androgens and progestogens, thus making its application for ‘real world’ samples very useful. In this work, different water-based samples from environmental origins were used to demonstrate the capacity of this new bioassay. Trials showed that most samples, with the exception of complex samples such as wastewater influent, can be assayed. © 2016 Elsevier B.V. All rights reserved.
⁎ Corresponding author at: Leibniz Institute of Plant Genetics and Crop Plant Research (IPK), Corrensstr. 3, D-06466 Gatersleben, Saxony-Anhalt, Germany. E-mail address: [email protected] (G. Kunze).
http://dx.doi.org/10.1016/j.scitotenv.2017.06.077 0048-9697/© 2016 Elsevier B.V. All rights reserved.
508
A. Chamas et al. / Science of the Total Environment 605–606 (2017) 507–513
1. Introduction Yeast-based bioassays, which can detect the biological activity of many compounds in the environment, play an important role in the field of bioanalytics. A significant proportion of these bioassays are dedicated to the detection of steroid hormones (Adeniran et al., 2015) as these can identify endocrine disrupting compounds (EDCs), compounds which can bind to hormone receptors and disturb the vertebrate endocrine system (Kabir et al., 2015; Casals-Casas and Desvergnes, 2011). There are yeast-based bioassays designed to detect estrogens (Kaiser et al., 2010; Sanseverino et al., 2008), progestogens (Viswanath et al., 2008; Chamas et al., 2015), glucocorticoids (Bovee et al., 2011; Pham et al., 2016), androgens (Bhattacharjee and Khurana, 2014; Gerlach et al., 2014) and more recently, for the simultaneous detection of several hormone classes (Chamas et al., 2017). Most of these bioassays combine the production of recombinant human hormone receptors and use enzymes or fluorescent reporter proteins to produce a quantitative signal. Contrary to standard analytical detection methods which give the exact composition of a sample, yeast-based bioassays quantify the total hormonal activity of a sample and therefore assess the biological impact a sample can have. Yeast-based bioassays are thus often used as a pre-screening procedure to give an overview of a samples hormonal activity before analytical detection methods are used to determine which compounds are present in the sample and identify potential EDCs. In an effort to combine the bioassay and analytical approaches, recent studies have introduced a bioautography detection method using thin-layer chromatography and cells from the yeast-based estrogen screen (YES) assay (Mueller et al., 2004; Spira et al., 2013). In these tests, an unknown sample is firstly separated on a thin-layer chromatography plate together with control molecules with known estrogenic activity. After separation, the plate is sprayed with an engineered yeast cell suspension producing a β-galactosidase when an estrogenic substance binds to the recombinant human estrogen receptor. The presence of this enzyme is then detected by spraying the plate again with a chromogenic or fluorescence-producing substrate. With this planarYES assay, compounds with estrogenic activity have been detected in various environmental samples (Klingelhöfer and Morlock, 2015; Buchinger et al., 2013), cosmetic products (Buchinger et al., 2013) and food samples (Klingelhöfer and Morlock, 2014). Detection of 17βestradiol in the femtogram range was reported in environment samples, thus enabling ultratrace level determination. A great advantage of TLC based assays over regular microtiter plate based assays is that prior separation of a complex sample during thin-layer chromatography will remove compounds that could be toxic to yeast. Another advantage of the planar-YES assay is that it requires less than 24 hours incubation to obtain a signal after chromatographic separation whereas the regular YES assay requires incubation for 3 to 5 days to obtain an unequivocal result (Routledge and Sumpter, 1996). To obtain sharp bands and very low detection limits, the latest optimizations of the planar-YES assay employs high performance thin layer chromatography plate coatings such as Reverse phase RP-18W (Klingelhöfer and Morlock, 2014) and fluorescence-producing substrate 4-methylumbelliferyl-b-D-galactopyranoside (MUG) (Buchinger et al., 2013). These improvements make the planar-YES assay of comparable performance to the established analytical methods, offering determination of a samples biological activity at low concentrations as well as identification of the compounds. However, the assay is more complex and more expensive than previous assays which may reduce the attractiveness of this analytical/biological detection method. In an effort to incease simplicity in preference to low sensitivity and a broad detection range, we designed a bioassay combining analytical separation with unmodified silica gel thin-layer chromatography plates and detection without addition of an enzyme substrate. To achieve the latter goal, fluorescent reporter proteins were used, obviating the need for substrate. Moreover, ideal EDC detection should not only detect estrogenic molecules but also other EDCs such as progestogens
and androgens. Therefore our aim was to design a bioassay able to determine in one analysis which endocrine activity is present in a sample and which compounds are responsible for it. This assay, to our knowledge, is the first to possess these characteristics and could be developed as an assay for other molecules. Modified Arxula adeninivorans strains were used as the biocomponent in the assay. These strains, described in a recently published research article (Chamas et al., 2017), produce a recombinant human progesterone, androgen or estrogen receptor, each in combination with the production of a fluorescent reporter protein, CFP, GFP or DsRed2 respectively. In the microtiter plate assay, these strains were mixed and could then, once incubated with a mixture of estrogens, androgens and progestogens, determine which particular hormone activity is present in the sample. It was then decided to test if this yeast strain mix can be combined with thin-layer chromatography and after three fluorescence scans of a plate, determine which hormone activities are present in a sample. To achieve this, the bioassay should realise these objectives: (1) all three A. adeninivorans strains should be able to interact with their target hormone on a thin-layer chromatography plate, (2) the fluorescence emitted from this interaction should be detectable with a plate scanner, (3) the three yeast strains, when mixed, should give easily distinguishable signals for each of the three hormonal types, (4) the development conditions for the assay should be uncomplicated. 2. Material and methods 2.1. Chemicals Estrone (E1), 17β-estradiol (E2), estriol (E3), progesterone (P), medroxyprogesterone acetate (MP), mifepristone (M), nandrolone (N), trenbolone (T), 5α-dihydrotestosterone (DHT) were all purchased from Sigma-Aldrich (Steinheim, Germany). All hormones were solubilized in ethanol at a stock concentration of 1 g/l. Acetone, dichloromethane and cyclohexane were all also purchased from Sigma-Aldrich (Steinheim, Germany). 2.2. Strains and cultivation conditions The yeast strains used in this work, (A. adeninivorans G1212/ YRC102-hPR-CFP, G1212/YRC102-hAR-GFP and G1212/YRC102-hERDsRed2), are described in a previous publication (Chamas et al., 2017). Each strain was cultivated for 24 h in yeast minimal media supplemented with glucose (Yeast Minimal Medium [YMM]-glucose) at 30 °C until the OD620 nm reached approximately 3. The cells were pelleted by centrifugation at 4000 × g for 10 min, the pellet was resuspended in an equal volume of fresh YMM-maltose and the three cell suspensions were mixed together to create A. adeninivorans G1212/YRC102-hHRfluo at a final total OD620 nm of 3. 2.3. Thin-layer chromatography procedure 100 × 200 mm polyester sheets precoated with 0.2 mm silica gel and without fluorescent markers (Macherey-Nagel, Düren, Germany) were used for all thin-layer chromatography experiments. For automated application of samples, the desired volumes were applied using an automatic TLC Sampler 4 (CAMAG, Muttenz, Switzerland) as 6 mm bands, 8 mm from the bottom of the plate. The application zones were allowed to dry at room temperature for 5 min and the plate was inserted in a TLC vertical development chamber containing dichloromethane:cyclohexane:acetone (60:35:5, v/v/v). Once the solvent migrated to within 85 mm from the bottom of the plate, the plate was removed from the developing chamber and allowed to dry at room temperature for 30 min. For manual sample application, the desired volume was applied with a glass capillary at 10 mm from the bottom of the plate. After drying for 5 min at room temperature, the plate was focused two times with methanol until the solvent
A. Chamas et al. / Science of the Total Environment 605–606 (2017) 507–513
reached 20 mm so that sharp bands formed before development with the solvent. Between each focusing, the plate was dried at 50 °C for 5 min to completely remove methanol. Vertical development occurred until the dichloromethane:cyclohexane:acetone solvent reached 85 mm from the bottom of the plate.
2.4. Chemical detection Developed plates were sprayed with 8% sulfuric acid in ethanol and incubated for 15 min at 110 °C. Absorbance at 310 nm and 500 nm of each track was then measured with a TLC Scanner 3 (CAMAG, Muttenz, Switzerland). The CAMAG-embedded software, Wincats was used to provide a graphical display of the absorbing zones and calculate the peak areas of the peaks.
2.5. Bioassay detection The developed plates were immersed in A. adeninivorans G1212/ YRC102-hHR-fluo for 5 s and the excess cell suspension was removed with a clean paper towel. The plate was then inserted in a closed plastic box with wet towel and the box was incubated for 18 h at 30 °C. The plate was dried at room temperature for 30 min and three fluorescence scans of each track were performed to determine CFP, GFP and DsRed2 fluorescence. The excitation wavelengths and emission filters were respectively 445 nm/K460 nm for CFP, 485 nm/K500 nm for GFP and 542 nm/K560 nm for DsRed2.
509
2.6. Wastewater samples The bioassay was trialed with environmental samples from seven different origins: stream (Saxony, Germany), source water (Brandenburg, Germany and Saxony, Germany), river (Saxony, Germany), lake (Berlin, Germany), tap water (Berlin, Germany), wastewater plant influent (North Rhine-Westphalia, Germany) and wastewater plant effluent (Rhineland-Palatinate, Germany). Precise locations of these sources can be provided upon request. Samples were not pretreated or modified prior to the assay. To assess the suitability of the combined TLC assay for environmental analysis, a mix of the three hormones E2, P and DHT (all at the concentration 1 μg/l) was added to double distilled water (positive control) and to all samples. Column-mediated extraction of the positive control and the samples was performed with an enrichment factor of 20. Each sample (5 μl volume) was then manually applied to the TLC plates for bioassay detection. After the three fluorescence scans, the area of the different peaks was calculated with Wincats to allow comparison between the positive control and the samples. 3. Results and discussion 3.1. Determination of minimum detectable amount of each hormone class Different quantities of the three hormones progesterone (P), 5αdihydrotestosterone (DHT) and 17β-estradiol (E2) solubilized in 5 μl ethanol, were applied separately to three thin-layer chromatography plates and submitted to biological detection with the three strains of A. adeninivorans without chromatographic separation. At the end of
Fig. 1. (A) Densitogram of a TLC plate incubated with A. adeninivorans G1212/YRC102-hER-DsRed2 where different amounts of E2 were applied. The fluorescence scan was performed at 542 nm (excitation) with a K560 nm filter. The relationship between peak area and amount of hormone applied to TLC plates and incubated with A. adeninivorans G1212/YRC102-hERDsRed2, G1212/YRC102-hAR-GFP and G1212/YRC102-hPR-CFP is shown in (B), (C) and (D), respectively.
510
A. Chamas et al. / Science of the Total Environment 605–606 (2017) 507–513
the 18 hours incubation, the three plates were scanned at the appropriate wavelength for the reporter protein concerned to obtain three different densitograms. An example of densitogram obtained when a TLC plate with different quantities of E2 was incubated with A. adeninivorans G1212/YRC102-hER-DsRed2 is visible in Fig. 1A. A clear peak is visible for amounts of hormone superior to 7.5 pg. After determining the area of the peak via the CAMAG software, Wincats, a diagram illustrating the relationship between the amount of applied E2 and the peak area was created and can be seen on Fig. 1B. Additionally, the same diagrams were created for the plate incubated with A. adeninivorans G1212/YRC102-hPR-CFP (Fig. 1C) and G1212/ YRC102-hAR-GFP (Fig. 1D). In all three cases, a clear increase in the peak area with increasing hormone amount following a 4-parameters logarithmic curve can be seen. By reproducing these experiments, it appeared that even though the minimal detectable amount of each hormone remained the same, significant differences between the peak areas for identical quantities of hormone were seen on different plates (data not shown). Thus, in all cases, similar 4-parameters logarithmic curves were seen but with different maximum values. To explain this phenomenon, one hypothesis is that thin-layer chromatography plates vary from batch to batch in terms of silica coating thickness and quality. As it was already shown that yeast actually grow on silica plate (Klingelhöfer and Morlock, 2014), these structural differences of the plate can lead to yeast growth variations and therefore to higher or lower fluorescence. From all experiments, minimal detectable amounts of 7.5 pg E2, 25 pg P and 25 pg DHT were detected, which correspond to initial concentrations applied to the plates of 1.5 μg/l, 5 μg/l and 5 μg/l, respectively. Even though these concentrations seem high when compared to the detection limits of microtiter plate based yeast assays, it is important to understand that more than 5 μl of an ethanol-based hormone dilution can be applied in the same track, thus allowing the utilization of lower dilutions than 1.5 μg/l E2, 5 μg/l progesterone and 5 μg/l DHT which were made for each hormone. Additionally, because ethanol evaporates quickly, successive applications of this hormone can be performed with intermediary drying phase, thus lowering again the bioassay detection limit. Another important point which can extracted from
Fig. 1A is that, even though the peaks get slightly broader with increasing E2 amount, this effect remains small when compared to the results obtained with a similar plate in the paper of Mueller et al. (2004). This indicates that the phenomenon responsible for peak diffusion in silicabased TLC plates without chromatographic separation may not be the long incubation with water-based cell suspension but rather the second incubation with the water-based substrate solution at higher temperature (37 °C). 3.2. Separation of estrogens, androgens and progestogens by TLC – chemical detection Thin-layer chromatography, due to the relatively small plates, allows the separation of only a limited number of compounds and this separation occurs on the basis of the compounds adsorption properties on the solid phase as well as its solubility in the mobile phase. These physical properties are frequently related to the atomic structure of the molecule and consequently, similar molecules will be harder to separate in a TLC format. Steroid hormones, all originating from the molecule cholesterol, have a similar structure and therefore several solvents were trialed to find an acceptable universal mobile phase for estrogens, progestogens and androgens. The best separation was obtained with a mix of dichloromethane, cyclohexane and acetone with the volume proportions 60:35:5. To show this separation, three estrogens (estrone (E1), E2 and estriol (E3)), three progestogens (P, medroxyprogesterone acetate (MP) and mifepristone (M)) and three androgens (DHT, nandrolone (N) and trenbolone (T)) as well as mixtures of these estrogens, progestogens and androgens were applied to a TLC plate; 360 ng of each compound per track. After separation, an non-specific chemical detection was performed by spraying 8% sulfuric acid and incubating the plate at 110 °C for 15 min. Strong acid is known to oxidize a very wide range of molecules and steroid hormones in particular. This oxidation produces a dark stain on the TLC plate (Bhawani et al., 2010) which can be scanned at 500 nm, with the exception of oxidized progesterone which is only visible under UV. The densitogram obtained after TLC plate scan is visible in Fig. 2. Each visible peak was then attributed to a
Fig. 2. Densitogram of a TLC plate where several hormones of different classes were applied and allowed to migrate in dichloromethane:cyclohexane:acetone solvent. Track 1: E1, track 2: E2, track 3: E3, track 4: mix of E1, E2 and E3, track 5: P, track 6: MP, track 7: M, track 8: mix of P, MP and M, track 9: DHT, track 10: N, track 11: T, track 12: mix of DHT, N and T. The scan was performed at 500 nm after chemical detection.
A. Chamas et al. / Science of the Total Environment 605–606 (2017) 507–513 Table 1 Retention factor of selected hormones determined with chemical detection. Hormone class
Hormone
Rf
Estrogens
Estrone 17β-Estradiol Estriol Progesterone Medroxyprogesterone acetate Mifepristone 5α-Dihydrotestosterone Nandrolone Trenbolone
0.49 0.26 0.03 0.53 0.51 0.28 0.37 0.24 0.01/0.23
Progestogens
Androgens
specific hormone and the retention factor of each compound was determined. These values can be found in Table 1. For progesterone, the plate was scanned at 310 nm to determine the retention factor (TLC scan not shown). A good separation of the three estrogens E1, E2 and E3 is visible with this mobile phase with retention factors separated by at least 0.2. P and MP also show similar retention factor values as would be expected from
511
the similarity of these compounds however the third progestogen, mifepristone, has a much lower retention factor. The three androgens show relatively good separation with the particularity that T presents two peaks, one with a retention factor of 0.01 and the second by 0.23, probably indicating that the commercial T consists in two isoforms with two different chromatographic properties. It is also important to note that the three most common steroid hormones - E2, P and DHT – show well separated retention factors. 3.3. Separation of estrogens, androgens and progestogens by TLC – fluorescent detection It has been shown that the three strains of A. adeninivorans G1212/ YRC102-hPR-CFP, G1212/YRC102-hAR-GFP and G1212/YRC102-hERDsRed2 can be simultaneously incubated with hormones from the three different classes in microtiter plate format (Chamas et al., 2017) and give three distinct signals. To test whether such incubation can also be applied to the combination assay, 100 pg of E2, P and DHT and a mix of these three hormones was applied to a TLC plate and separated with dichloromethane:cyclohexane:acetone mobile phase. After
Fig. 3. Densitogram of a TLC plate allowed to migrate in dichloromethane:cyclohexane:acetone solvent where E2 (violet line), P (blue line) and DHT (green line) as well as a mix of these hormones (yellow line) were applied. Fluorescence scans at the wavelength 445 nm/K460 nm for CFP (A), 485 nm/K500 nm for GFP (B) and 542 nm/K560 nm for DsRed2 (C) were performed after incubation of the plate with A. adeninivorans G1212/YRC102-hHR-fluo.
512
A. Chamas et al. / Science of the Total Environment 605–606 (2017) 507–513
development, the plate was dipped into a suspension of A. adeninivorans G1212/YRC102-hPR-CFP, G1212/YRC102-hAR-GFP and G1212/ YRC102-hER-DsRed2. After 18 hours incubation the plate was scanned at three different wavelengths to detect CFP, GFP and DsRed2 fluorescence as described in Material and Methods. The densitograms obtained for these three scans can be seen in Fig. 3. For each densitogram, the fluorescence of the four tracks were recorded and plotted on the same diagram with blue, green, violet and yellow curves corresponding respectively to the tracks where P, DHT, E2 and a mixture of the three were applied. In comparison to the detection without chromatographic development, the shape of these peaks is more disturbed and the baseline is not completely flat. Nevertheless, peaks are present in all three scans and by comparing the Rfs of these peaks with the Rfs determined in the preceding section, all three hormones can be successfully detected. When the plate was scanned for GFP fluorescence (Fig. 3.B), the peak of E2 is clearly visible in the track where E2 only was applied and in the track with a mix of the three hormones. This could indicate that the A. adeninivorans G1212/YRC102-hAR-GFP cells can detect E2, but previous work showed that the androgen receptors affinity for E2 is very low (Gerlach et al., 2014; Scippo et al., 2002). To confirm this, an additional strain producing the androgen receptor and possessing CFP as reporter protein was created and incubated with a TLC plate where E2 was applied. This experiment showed no E2 peak at all (data not shown). To explain the GFP fluorescence observed at the Rf of E2, the most plausible answer is found in the fluorescence detector. The detector of the CAMAG scanner 3 is preceded by fluorescence filters which will cut all light of wavelengths shorter than the desired wavelength, thus only allowing detection of the emitted light and not the excitation light. In the case of the GFP fluorescence detection, in order to excite A. adeninivorans G1212/YRC102-hAR-GFP cells at 485 nm and measure fluorescence at 520 nm, the K500 filter was used to remove emission light with wavelengths below 500 nm before it strikes the detector. It was shown by Bevis and Glick (2002) that DsRed possesses two excitation peaks, one small at 488 nm and a larger one at 558 nm, and one emission peak at 583 nm. Therefore, the 485 nm excitation will also induce dsRED fluorescence through the small excitation peak and the emitted light from this fluorescent protein will not be removed by the K500 filter. To overcome this phenomenon, two main strategies can
be followed. Firstly the DNA sequence of the DsRed2 protein can be genetically modified to remove the excitation peak at 488 nm. Secondly, the utilization of band filters in the Scanner detector could allow more precise selection of which emitted wavelength can be observed. The use of such a filter is not possible with the CAMAG Scanner 3 but other models offer this possibility. 3.4. Application of wastewater samples – fluorescent detection One of the applications of the combined bioassay is to detect hormone activity in environmental samples. However it has already been shown that such samples generally have relatively low concentrations of steroid hormones, mostly in the ng/l range for estrogens for example. Consequently, it was decided to use spiked samples as described in the Materials and methods section. The same mix of E2, P and DHT was added to either the environmental samples or double distilled water (the positive control) and, after extraction, 5 μl of the eluted volume was applied to a TLC plate. After the three scans for CFP, GFP or DsRed2 fluorescence, the peak areas for each hormone in each sample and in the positive control were determined. To compare the hormone activity of each sample, a simple normalization was performed by dividing the peak area of one hormone in a sample by the peak area of the same hormone in the positive control. A value of 1 indicates that the same hormone concentration was detected in the sample and in the positive control. This allows assessment of whether or not the sample matrix has an effect on the detection of hormones. The same test was performed on two different TLC plates on different days and the results of these two experiments were used for statistical analysis. The relative hormone activity of each sample for the three hormones is presented in the three diagrams in Fig. 4 with error bars representing the 95% confidency interval of the measurement. The samples, source 1, river 1, river 2, tap water and wastewater effluent show values around 1 for each three hormones, thus indicating that the spiked concentration of hormone was producing similar peaks in the positive control and in the samples. The sample stream presents ratios of 1 for E2 and DHT but for P, the ratio is slightly higher. In the case of samples source 2 and wastewater influent 2, all three hormones have been detected but all ratio are inferior to 1, indicating a smaller area of the hormone
Fig. 4. Ratio values of androgens (blue bars), estrogens (red bars) and progestogens (green bars) in different environmental samples. Ratio value is calculated by dividing the peak area of a given hormone in a spiked sample by the peak area of the same hormone in the positive control. Error bars indicate the 95% confidence interval of the ratio.
A. Chamas et al. / Science of the Total Environment 605–606 (2017) 507–513
peak in those samples if compared to the positive control. Finally, the lake and wastewater influent 1 samples show high androgens and estrogen ratio but with large error bars. This wide 95% confidence interval shows that, for these two types of samples, the reliability of the measurement is low, thus indicating that the detection of endocrine active compounds in such complex samples with the described assay has to be improved. The reason for this low reliability may be that additional compounds which are not hormones but which are extracted simultaneously, interfere with the yeast metabolism or the fluorescence molecule. The same limitations were also seen with a yeast assay in microtiter plate format where environmental samples with complex matrices were found to be cytotoxic in many cases (Gehrmann et al., 2016). The fact that detecting estrogens in wastewater treatment plants influents is possible has been demonstrated by Klingelhöfer and Morlock (2015) and therefore the limitations observed in the present study regarding this type of sample may have to be overcome with an improved extraction process. Taken together, these results indicate that for samples with relatively simple composition and good purity like tap water or wastewater effluent, the separation and determination of the three steroid hormones is possible. However, for more complicated samples such as wastewater treatment plant influent, the bioassay protocol has to be optimized. 4. Conclusions The bioassay developed here enables the separation and a rapid identification of hormones of different classes within one day. The use of A. adeninivorans G1212/YRC102-hHR-fluo biocomponent allows simultaneous detection of estrogens, androgens and progestogens on one plate with only a fluorescence plate reader to perform the assay. Because the potential applications of this combined assay are in environmental science, a field where fully-equipped laboratories with analytical measurement devices are not always available, special attention was given to keep the assay as simple as possible. Simple TLC plates with silica matrix, common solvents and manual sample application are all that is required for the standard procedure which allowed the detection of well-separated peaks. The use of fluorescence reporter proteins removes the need of a second incubation step, which is necessary in the case of enzyme-mediated detection, and allows the detection of the signal in less than 24 h. However the assays limit of detection needs to be reduced in order to identify and quantify EDCs present in environmental samples. Future work will therefore focus on identifying or developing a simple extraction process for complex samples and development of yeast strains with improved fluorescence reporters. With this bioassay, EDCs can also be identified by simply comparing retention factors of active compounds with the retention factors of known hormones. In future work, establishment of a method to remove active compounds by scratching from the thin-layer chromatography plate for subsequent analytic detection by HPLC could provide a method to corroborate the identity of molecules identified by the assay and also potentially allow the discovery of new EDCs. Acknowledgements The work was supported by grant from the German Federal Ministry of Economics and Technology (Bundesministerium für Wirtschaft und Technologie, BMWi) within its program “Zentrales Innovationsprogramm
513
Mittelstand, ZIM” (Grant Numbers KF2131626SB4, KF2715706SB4, KF2025438SB4). References Adeniran, A., Sherer, M., Tyo, K.E.J., 2015. Yeast-based biosensors: design and applications. FEMS Yeast Res. 15, 1–15. Bevis, B.J., Glick, B.S., 2002. Rapidly maturing variants of the Discosoma red fluorescent protein (DsRed). Nat. Biotechnol. 20, 83–87. Bhattacharjee, G.B., Khurana, S.M.P., 2014. In vitro reporter assays for screening of chemicals that disrupt androgen signalling. J. Toxicol. 2014, ID 701752. http://dx. doi.org/10.1155/2014/701752. Bhawani, S.A., Sulaiman, O., Hashim, R., Mohamad Ibrahim, M.N., 2010. Thin-layer chromatographic analysis of steroids: a review. Trop. J. Pharm. Res. 9, 301–313. Bovee, T.F., Helsdingen, R.J., Hamers, A.R., Brouwer, B.A., Nielen, M.W., 2011. Recombinant cell assay for the detection of (gluco)corticosteroids and endocrine-disrupting potencies of several environmental PCB contaminants. Anal. Bioanal. Chem. 401, 873–882. Buchinger, S., Spira, D., Bröder, K., Schlüsener, M., Ternes, T., Reifferscheid, G., 2013. Direct coupling of thin-layer chromatography with a bioassay for the detection of estrogenic compounds, application for effect-directed analysis. Anal. Chem. 85, 7248–7256. Casals-Casas, C., Desvergnes, B., 2011. Endocrine disruptors: from endocrine to metabolic disruption. Annu. Rev. Physiol. 73, 135–162. Chamas, A., Nieter, A., Pham, H.T.M., Giersberg, M., Hettwer, K., Uhlig, S., Simon, K., Baronian, K., Kunze, G., 2015. Development of a recombinant Arxula adeninivorans cell bioassay for the detection of molecules with progesterone activity in wastewater. Anal. Bioanal. Chem. 407, 8109–8120. Chamas, A., Pham, H.T.M., Jähne, M., Hettwer, K., Uhlig, S., Simon, K., Einspanier, A., Baronian, K., Kunze, G., 2017. Simultaneous detection of three sex steroid hormone classes using a novel yeast-based biosensor. Biotechnol. Bioeng. 114 (2017), 1539–1549. Gehrmann, L., Bielak, H., Behr, M., Itzel, F., Lyko, S., Simon, A., Kunze, G., Dopp, E., Wagner, M., Tuerk, J., 2016. (Anti-)estrogenic and (anti-)androgenic effects in wastewater during advanced treatment: comparison of three in vitro bioassays. Environ. Sci. Pollut. Res. Int. 1–11. Gerlach, T.K., Kaiser, C., Körner, M., Hettwer, K., Uhlig, S., Simon, K., Baronian, K., Kunze, G., 2014. Development and assessment of a novel Arxula adeninivorans androgen screen (A-YAS) assay and its application in analysis of cattle urine. Sci. Total Environ. 490, 1073–1081. Kabir, E.R., Rahman, M.S., Rahman, I., 2015. A review on endocrine disruptors and their possible impacts on human health. Environ. Toxicol. Pharmacol. 40, 241–258. Kaiser, C., Uhlig, S., Gerlach, T., Körner, M., Simon, K., Kunath, K., Florschütz, K., Baronian, K., Kunze, G., 2010. Evaluation and validation of a novel Arxula adeninivorans estrogen screen (nAES) assay and its application in analysis of wastewater, seawater, brackish water and urine. Sci. Total Environ. 408, 6017–6026. Klingelhöfer, I., Morlock, G.E., 2014. Sharp-bounded zones link to the effect in planar chromatography-bioassay-mass spectrometry. J. Chromatogr. A 1360, 288–295. Klingelhöfer, I., Morlock, G.E., 2015. Bioprofiling of surface/wastewater and bioquantitation of discovered endocrine-active compounds by streamlined direct bioautography. Anal. Chem. 87, 11098–11104. Mueller, M.B., Dausend, C., Weins, C., Frimmel, F.H., 2004. A new bioautographic screening method for the detection of estrogenic compounds. Chromatographia 60, 207–211. Pham, H.T.M., Chamas, A., Nieter, A., Giersberg, M., Rutten, T., Gehrmann, L., Hettwer, K., Tuerk, J., Uhlig, S., Simon, K., Baronian, K., Kunze, G., 2016. Determination of glucocorticoids using photometric (A-YGS) and spectrofluorometric (A-YGFS) bioassays based on modified Arxula adeninivorans cells: applications in environmental analysis. Sensors Actuators B Chem. 223, 540–549. Routledge, E.J., Sumpter, J.P., 1996. Estrogenic activity of surfactants and some of their degradation products assessed using a recombinant yeast screen. Environ. Toxicol. Chem. 15, 241–248. Sanseverino, J., Elridge, M.L., Layton, A.C., Easter, J.P., Yarbrough, J., Schultz, T.W., Sayler, G.S., 2008. Screening of potentially hormonally active chemicals using bioluminescent yeast bioreporters. Toxicol. Sci. 107, 122–134. Scippo, M.L., Van de Weerdt, C., Willemsen, P., Francois, J.M., Rentier-Delrue, F., Müller, M., Martial, J.A., Maghuin-Rogister, G., 2002. Detection of illegal growth promoters in biological samples using receptor binding assays. Anal. Chim. Acta 473, 135–141. Spira, D., Reifferscheid, G., Buchinger, S., 2013. Combination of high-performance thinlayer chromatography with a specific bioassay – a tool for effect-directed analysis. J. Planar Chromatogr.–Mod. TLC 26, 395–401. Viswanath, G., Halder, S., Divya, G., Majumder, C.B., Roy, P., 2008. Detection of potential (anti)progestagenic endocrine disruptors using a recombinant human progesterone receptor binding and transactivation assay. Cell. Mol. Endocrinol. 295, 1–9.