Screening for Phytoestrogens using a Cell-based Estrogen Receptor β Reporter Assay
Summary
We have optimized a commercially available estrogen receptor β reporter assay for screening human and nonhuman primate foods for estrogenic activity. We validated this assay by showing that the known estrogenic human food soy registers high, while other foods show no activity.
Abstract
Plants are a source of food for many animals, and they can produce thousands of chemicals. Some of these compounds affect physiological processes in the vertebrates that consume them, such as endocrine function. Phytoestrogens, the most well studied endocrine-active phytochemicals, directly interact with the hypothalamo-pituitary gonadal axis of the vertebrate endocrine system. Here we present the novel use of a cell-based assay to screen plant extracts for the presence of compounds that have estrogenic biological activity. This assay uses mammalian cells engineered to highly express estrogen receptor beta (ERβ) and that have been transfected with a luciferase gene. Exposure to compounds with estrogenic activity results in the cells producing light. This assay is a reliable and simple way to test for biological estrogenic activity. It has several improvements over transient transfection assays, most notably, ease of use, the stability of the cells, and the sensitivity of the assay.
Introduction
Plants are a necessary source of food for many animals, providing calories and nutrients critical to survival, reproduction, growth, development, and behavior1. Plants produce thousands of chemicals, many as adaptations for their own growth, stomatic maintenance, and reproduction. Other compounds, deemed plant secondary metabolites (PSMs), have functions that are less clear, though some are toxic and likely used as a defense against herbivory and parasitism (e.g., alkaloids, tannins)2,3. Some of these chemicals have the ability to affect long term physiological processes in animals, such as endocrine functioning, although why these endocrine-active phytochemicals interact with the vertebrate endocrine system is still unclear2,4.
Phytoestrogens, the most well studied endocrine-active phytochemicals, are polyphenolic PSMs that structurally and functionally mimic estrogens, directly interacting with the hypothalomo-pituitary gonadal axis of the vertebrate endocrine system5. Ingestion of phytoestrogens in the human diet is associated with protection against some cancers, heart disease, and menopausal symptoms, though other effects include fertility problems. In fact, the physiological effects of these compounds were discovered in the 1940s when infertility in sheep was attributed to their grazing on phytoestrogen-rich clover (Trifolium subterrareum)6. When ingested, phytoestrogens can pass into cells and mimic the effects of estrogen. While phytoestrogens had negative effects on sheep fertility, the relationship between phytoestrogens and physiology is not simple. Like sheep, southern white rhinoceros display sensitivity to estrogenic compounds in feed derived from high quantities of soy and alfalfa. Daughters of females fed this diet during pregnancy are less likely to reproduce7. However, other studies have shown that phytoestrogens may have positive effects as well, including maturation of ovarian follicles in older mice8, prevention of certain cancers, antioxidant activity, and antiproliferative effects9.
The breadth of effects of phytoestrogens are not surprising given that estrogens affect a wide array of biological functions, including growth, development, and regulation of the reproductive and central nervous systems10. Although there are many mechanisms of action, phytoestrogens often have the ability to modify, enhance, or disrupt estrogen signaling through their ability to act as ligands for the intranuclear estrogen receptors alpha and beta (ERα and ERβ). Many phytoestrogens have a phenolic ring structure similar to estrogens that allows them to bind estrogen receptors. Those with agonistic estrogenic activity function like estrogen, forming an activated ER-ligand complex that can dimerize and bind to an estrogen response element (ERE) and trigger gene transcription11. Thus, estrogens and phytoestrogens regulate cell activity and system functions through their actions as transcription factors.
Here we present the novel use of a cell-based assay to screen plant extracts for the presence of compounds that have estrogenic biological activity. This assay uses Chinese hamster ovary CHO cells engineered to highly express ERβ, which have been transfected with the firefly (Photinus pyralis) luciferase gene linked to an ERE promoter12. When estrogenic compounds are present, they bind to the ER, dimerize, and bind to the ERE, leading to transcription of the luciferase gene. Upon addition of a substrate solution, the luciferase catalyzes a reaction leading to photon emission. Therefore, positive samples produce light and negative samples do not.
This commercially available assay eliminates the need for laboratories to transfect the mammalian cells with the reporter gene and estrogen receptor13,14, which was unstable and variable in efficacy. The assay provides a stable transfection platform that allows for quickly and simply determining whether a plant has estrogenic activity via receptor binding.
We test the hypothesis that soybeans have higher estrogenic activity than all other foods given their known concentrations of estrogenic isoflavones15 using human foods from local grocers.
Protocol
1. Preparation of plant materials
- Freeze dry plant items that were collected fresh using a lyophilizer.
- To protect samples from light, cover chambers with aluminum foil during drying process.
- To ensure that samples are completely dry, lyophilize until chambers no longer feel cold to touch and plant materials no longer lose mass when weighed.
- Store dried plants in sterile low residue bags in absence of light until grinding.
- Finely grind samples using a grinding mill with 0.85 mm mesh screen.
- Store the ground samples in the bags in the absence of light until extraction.
2. Extraction of plant secondary metabolites
- To extract the secondary plant metabolites, use a ratio of 1 g of dried sample to 10 mL of HPLC grade methanol.
- Weigh sample on an analytical balance and add it to an appropriately sized Erlenmeyer flask (125 – 250 mL). Then add appropriate volume of methanol. Record the mass of sample extracted.
- Cover the plant-methanol solution with aluminum foil, and then set to rotate at 100 rpm speed at room temperature (RT) for 3 days on an orbital shaker, allowing the potentially estrogenic compounds to dissolve into the methanol.
- Decant the supernatant into a drip filtration system using filter paper (125 mm).
- Using a rotary evaporator, dry the plant extract until the sample is thickened, but pourable, in a 300 mL round-bottom flask. Pour sample into a 50 mL round-bottom flask, rinsing the large flask with a small amount of methanol. Continue to dry the sample in the small flask until the methanol is completely evaporated.
- Weigh the sample residue using an analytical balance. Record residue mass.
- Dissolve the plant extract in dimethyl sulfoxide (DMSO) at a concentration of 0.1 g of extract to 2 mL of DMSO. Vortex until homogenized.
- Store the plant extract-DMSO solution at 4 °C in amber glass vials until the assay.
CAUTION: Plants can produce unknown biologically active chemicals, and DMSO is a vehicle that can transport them across cell membranes. Use appropriate personal protective equipment and care when handling these samples.
3. Human estrogen receptor β transfection assay12
NOTE: Aseptic technique and a laminar flow hood is required for Day 1 of the assay protocol.
- Prepare dilutions of 17β-Estradiol for the standard curve.
- Transfer the Cell Recovery Medium and Compound Screening Medium (CSM) from the freezer storage and thaw in a 37 °C water bath.
- Label microcentrifuge tubes Intermediate 1 and 2 (INT1, INT2) and 1-8.
- Fill INT1 with 995 μL of CSM, INT2 with 615 μL of CSM, tube 1 with 900 μL of CSM, and tubes 2-8 with 600 μL of CSM. Set tube 8 aside.
- Transfer 5 μL of 100 μM 17β-Estradiol Stock into INT1. Discard the tip. Vortex.
- Before each transfer, rinse pipette 3 times, and then transfer 10 μL from INT1 into INT2. Discard the tip.
- Rinse pipette 3 times, and then transfer 100 μL from INT2 into tube 1. Discard tip. Transfer 300 μL from tube 1 into tube 2. Repeat for tubes 3 through 7. Discard 300 μL from tube 7 into waste container. Tube 8 is a Zero and does not receive estradiol. Final concentrations of plated standards are: 400, 133.3, 44.44, 14.815, 4.938, 1.646, 0.5487, and 0 pM estradiol.
- Prepare sample compounds.
- Vortex samples.
- Take 4 μL of each plant sample in DMSO and add to 496 μL of CSM to yield a 0.8% DMSO solution.
- Rapidly thaw Reporter Cells.
- Retrieve the tube of Cell Recovery Medium from the 37 °C water bath. Disinfect the outside surface using 70% ethanol.
- Retrieve Reporter Cells from -80 °C storage and thaw by transferring 10 mL of the pre-warmed CRM into the tube of frozen cells.
- Close the tube of Reporter Cells and transfer to a 37°C water bath for 5-10 min.
- Retrieve the tube of Reporter Cell Suspension from the water bath. Invert the tube of cells several times gently to break up aggregates of cells and produce a homogenous suspension. Clean the surface of the tube with 70% ethanol.
- Assay plating
- Dispense 100 μL of the Reporter Cell Suspension into each well using a multichannel pipette.
- Dispense 100 μL of samples in triplicate into appropriate assay wells.
- Transfer the plate into a 37 °C, humidified 5% CO2 incubator for 22-24 h.
- Thaw Detection Substrate and Detection Buffer in a dark refrigerator overnight to prepare for Day 2.
- Just prior to the end of the plate incubation, remove Detection Substrate and Detection Buffer from refrigerator and place in low light area until equilibrated to RT. Once at RT, invert each tube gently several times to thoroughly mix solutions.
- Immediately before the incubation is complete, pour the entire contents of the Detection Buffer into the tube of Detection Substrate to create Luciferase Detection Reagent. Mix gently so as not to produce foam.
- Once the incubation is complete, invert the plate to discard content into an appropriate waste container. Gently tap the plate on a clean absorbent paper towel to remove the last droplets from the wells.
- Add 100 μL of the Luciferase Detection Reagent to each well. Allow the assay plate to rest at RT for 15 min. Do not shake the plate.
- Quantify luminescence using a 96-well plate-reading luminometer.
Representative Results
Twenty-two extracts of fruits and vegetables commonly found in human diets were screened for the presence of estrogenic compounds. A variety of foods were assayed, including legumes, such as soybeans, snow peas, and snap peas, as the pea family is a known source of phytoestrogens16, as well as figs, dates, corn, carrots, apples, bananas, strawberries, tomato, kale, and cabbage. Endocrine disrupting compounds are found in common substances (e.g., plastics and pesticides) and some are biologically active through ERs17. When possible, both organic and nonorganically grown items were assayed to account for the possibility that pesticides with estrogenic activity could have affected the results.
Each plant food item was plated in triplicate and the luminometer reported each well’s activity in Relative Light Units (RLUs). Background levels of RLUs are determined in the standard curve with Standard 8, the zero concentration, and used for reference. The fold activation value, which is the multiplier above the RLU for the Zero point on the curve, is calculated by the equation:
Fold Activation = Unknown (RLU) ÷ Standard 8 (RLU)
For interpretive purposes, estrogenic activity is presented in an ordinal, qualitative manner of High, Med, Low, or No Activity. High levels of activity register above the Standard 4 fold activation value. Medium falls between Standard 5 and Standard 4, and Low values are between Standard 6 and Standard 5. Any samples with fold activation values below Standard 7 are considered No Activity. Referring to Table 1, soybeans, both organic and non-organic, screened at high levels of activity, while all other fruit and vegetable items registered no activity. Comparing soybean results to the standard curve (Figure 1), shows that, whether grown organically or not, they score high off the curve for estradiol activity levels at this concentration. Soybean extract, a known potent source of the isoflavones daidzein and genistein9, was further used to determine the dilution yielding a 50% signal to the maximum (Figure 2). This extract requires 422 times more dilution to produce half the signal of our standard dilution protocol.
| Produce Item | Organic/ Non-organic | Relative Light Units (Lum) | Fold Activation | Fold Activation (Mean) | Phytoestrogen Activity |
| Soybeans | Organic | 1687 | 29.016 | 31.06 | High |
| 2023 | 34.796 | ||||
| 1706 | 29.353 | ||||
| Soybeans | Non-organic | 2041 | 35.106 | 32.05 | High |
| 1956 | 33.647 | ||||
| 1593 | 27.399 | ||||
| Snow Peas | Non-organic | 53 | 0.919 | 0.92 | No Activity |
| 59 | 1.015 | ||||
| 49 | 0.836 | ||||
| Snap Peas | Non-organic | 66 | 1.142 | 1.21 | No Activity |
| 60 | 1.032 | ||||
| 85 | 1.462 | ||||
| Corn | Non-organic | 29 | 0.502 | 0.53 | No Activity |
| 30 | 0.513 | ||||
| 33 | 0.575 | ||||
| Strawberry | Non-organic | 35 | 0.609 | 0.77 | No Activity |
| 47 | 0.808 | ||||
| 51 | 0.884 | ||||
| Strawberry | Organic | 56 | 0.956 | 0.88 | No Activity |
| 59 | 1.015 | ||||
| 39 | 0.678 | ||||
| Banana | Organic | 32 | 0.544 | 0.52 | No Activity |
| 28 | 0.489 | ||||
| 31 | 0.533 | ||||
| Banana | Non-organic | 33 | 0.564 | 0.60 | No Activity |
| 41 | 0.712 | ||||
| 31 | 0.533 | ||||
| Plantain | Non-organic | 37 | 0.64 | 0.70 | No Activity |
| 39 | 0.667 | ||||
| 47 | 0.805 | ||||
| Kale | Organic | 26 | 0.447 | 0.47 | No Activity |
| 26 | 0.444 | ||||
| 30 | 0.519 | ||||
| Kale | Non-organic | 40 | 0.685 | 0.63 | No Activity |
| 28 | 0.485 | ||||
| 42 | 0.719 | ||||
| Cabbage | Organic | 33 | 0.568 | 0.54 | No Activity |
| 27 | 0.468 | ||||
| 34 | 0.588 | ||||
| Cabbage | Non-organic | 44 | 0.757 | 0.66 | No Activity |
| 34 | 0.585 | ||||
| 36 | 0.626 | ||||
| Apple | Organic | 30 | 0.523 | 0.49 | No Activity |
| 25 | 0.437 | ||||
| 30 | 0.509 | ||||
| Apple | Non-organic | 41 | 0.705 | 0.62 | No Activity |
| 31 | 0.53 | ||||
| 37 | 0.63 | ||||
| Tomato | Organic | 51 | 0.874 | 0.87 | No Activity |
| 57 | 0.974 | ||||
| 44 | 0.76 | ||||
| Tomato | Non-organic | 61 | 1.056 | 1.19 | No Activity |
| 81 | 1.386 | ||||
| 66 | 1.128 | ||||
| Carrot | Organic | 33 | 0.575 | 0.51 | No Activity |
| 33 | 0.561 | ||||
| 22 | 0.382 | ||||
| Carrot | Non-organic | 31 | 0.53 | 0.52 | No Activity |
| 21 | 0.365 | ||||
| 38 | 0.657 | ||||
| Fig | Non-organic | 29 | 0.506 | 0.61 | No Activity |
| 42 | 0.716 | ||||
| 36 | 0.619 | ||||
| Dates | Non-organic | 29 | 0.495 | 0.59 | No Activity |
| 39 | 0.667 | ||||
| 35 | 0.602 |
Table 1. Representative results of the ERβ Reporter Assay System for screening of fruit and vegetable items for phytoestrogen activity. Positive activity is indicated by High, Med, Low, or No Activity.
Figure 1. Serial dilution of 17β-Estradiol standard (Standard 1 through 8 concentrations = 400, 133.3, 44.44, 14.815, 4.938, 1.646, 0.5487, and 0 pM, respectively) using the ERβ Reporter Assay System. Please click here to view a larger version of this figure.
Figure 2. The ERβ Reporter Assay using a serial dilution of soybean extract to determine the dilution that yielded a signal-to-background ratio that is 50% of the maximum signal. From the standard extraction method dissolving the plant extract in dimethyl sulfoxide (DMSO) at a concentration of 0.1 g of extract to 2 mL of DMSO, soybean has to be diluted 422 times to elicit a signal 50% of the maximum response. Please click here to view a larger version of this figure.
Discussion
The ERβ reporter assay developed to individually screen pharmaceutical agents is also suitable for screening plant foods for phytoestrogens biologically active through the ERβ. Important considerations in the protocol include treating the plant samples with care: fresh plant material needs to be dried swiftly to prevent molding or other biological degradation, and it needs to be kept away from light to prevent photolysis of the compounds18. The assay protocol12 provided by the manufacturer is clear and needs very few modifications for screening purposes. The standard curve suggested by the manufacturer has been modified in this protocol to increase the number of points that fall in the exponential range of the curve (Figure 1), while preserving the top and bottom plateaus. It is possible to use this assay for quantitative analysis, but our purpose is to associate plants with high activity to biological effects, food choice, and other behaviors in the animals that consume them.
To further illustrate the effectiveness of the extraction and assay we included a dose response curve with soybean extract (Figure 2) and determined that given the potency of the normal extraction protocol, soy must be diluted extensively before the signal drops to 50% maximum. This highlights the fact that at high concentrations of phytoestrogens the signal plateaus at a stable maximum signal. At very low concentrations the signal may not be strong enough to be distinguished from background. It is important to work with high concentrations of extracts, in order to detect phytoestrogens present in low amounts in a sample, minimizing false negatives. Initially the laboratory used a greater volume of DMSO relative to the plant residue from the methanol extraction (i.e., 10 mL of DMSO to 0.1 g of plant residue). The samples were too dilute to induce a strong luminescence in positive samples. Due to a maximum DMSO percentage for reporter cell viability and volume constraints within the wells on the plate, sample extract concentration should be optimized when adding DMSO to the plant residues. A positive control such as soy should be included on every plate, to confirm that cells are viable and capable of luminescence, and that the extract concentration is sufficient to elicit a response.
This assay detects compounds that bind to ERβ, but not all phytoestrogens have the same mechanism of action. This assay protocol can be modified by incubating the cells with a combination of estradiol and the plant compounds to detect if there is antiestrogen activity in a sample9,12. Estradiol has great affinity to ER, so the presence of phytoestrogens may have antiestrogenic biological activity in the presence of estradiol by blocking the receptors, which reduces the response to estrogens. Antiestrogenic activity would be detected by a reduction in total activation with increasing concentration of plant extract. This assay will not detect other methods of action, such as binding to membrane-bound ERs19. Furthermore, some phytoestrogens are not biologically active until they have been metabolized by gut microbes20. It is possible that some plants that have no or low estrogenic activity in their unmetabolized state have higher estrogenic activity post-metabolization that this assay would not detect.
The ERβ reporter assay has been chosen to exemplify the screening of phytoestrogens for activity in plants because phytoestrogens compete for binding with estradiol more strongly to ERβ than they do to ERα21. Screening for ERα activity is possible through a similar assay, wherein the cells are transfected with the ERα gene rather than ERβ.
Following a positive screening for active phytoestrogens, the active compounds can be identified with chromatography methods. Indeed, at that point the isolated compounds can be tested using this assay and the half maximal effective concentrations (EC50) can be determined using a dilution series as a measure of potency of the compound.
This assay is a reliable and simple way to test for biological estrogenic activity, keeping in mind its limitations in the breadth of mechanisms of estrogenic activity. It has several improvements over transient transfection assays, most notably ease of use, the stability of the cells, and the sensitivity of the assay.
Little is known about the prevalence of phytoestrogens in wild plant foods consumed by humans or wild animals22, but studies show that exposure to estrogenic PSMs in diet can have long lasting effects23. Having a simple robust assay that detects these compounds, in conjunction with studies assessing amounts eaten and when they are eaten, is a powerful step in determining the function of including estrogenic foods in the diet and the effects of these compounds on physiological systems.
Divulgations
The authors have nothing to disclose.
Acknowledgements
Authors are thankful to Dale Leitman for initial training in use of transient transfection assays to determine estrogenic activity of primate plant foods. Thanks to Bradford Westrich and C. Eric Johnson for helping to set up laboratory equipment and training students in extraction methods. Finally, thank you to Indiana University for funding this research.
Materials
| 1000 µL pipette | |||
| 20 µL pipette | |||
| 200 µL pipette | |||
| 37 ℃ water bath | |||
| 37 ℃, humidified 5% CO2 incubator | |||
| 70% ethanol | |||
| analytical balance | |||
| cell culture-rated laminar flow hood | |||
| dimethyl sulfoxide | |||
| disposable media basin, sterile | |||
| drip filtration system | |||
| Erlenmeyer flasks | 125 mL and 250 mL | ||
| HPLC grade methanol | |||
| Human ERβ Reporter Assay System, 1 x 96-well format assays | Indigo Biosciences | IB00411 | Assay kit – analyzes 24 samples plus standard curve |
| lyophilizer | |||
| multi-channel pipette | |||
| orbital shaker | |||
| plate-reading luminometer | ex. Bioteck Synergy HTX | ||
| rotory evaporator | |||
| round bottom flasks | 50 mL and 300 mL | ||
| sterile microcentrifuge tubes or sterile multi-channel media basins | |||
| sterile tips | 200 µL and 1000 µL | ||
| Whatman grade 1 paper | |||
| whirl-pak bags | sterile polyethylene bags |
References
- Wasserman, M. D., et al. Estrogenic plant consumption predicts red colobus monkey (Procolobus rufomitratus) hormonal state and behavior. Hormones and Behavior. 62 (5), 553-562 (2012).
- Wasserman, M. D., Milton, K., Chapman, C. A. The roles of phytoestrogens in primate ecology and evolution. International Journal of Primatology. 34 (5), 861-878 (2013).
- DeGabriel, J. L., Moore, B. D., Foley, W. J., Johnson, C. N. The effects of plant defensive chemistry on nutrient availability predict reproductive success in a mammal. Ecology. 90 (3), 711-719 (2009).
- Wasserman, M. D., Steiniche, T., Després-Einspenner, M. -. L., Lambert, J. E., Rothman, J. M. . Primate Diet & Nutrition. , (2020).
- Benavidez, K. M., Chapman, C. A., Leitman, D. C., Harris, T. R., Wasserman, M. D. Intergroup variation in oestrogenic plant consumption by black-and-white colobus monkeys. African Journal of Ecology. , (2019).
- Bennetts, H. W., Underwood, E. J., Shier, F. L. A specific breeding problem of sheep on subterranean clover pastures in Western Australia. Australian Veterinary Journal. 22 (1), 2-12 (1946).
- Tubbs, C. W., et al. Estrogenicity of captive southern white rhinoceros diets and their association with fertility. General and Comparative Endocrinology. 238, 32-38 (2016).
- Shen, M., et al. Observation of the influences of diosgenin on aging ovarian reserve and function in a mouse model. European Journal of Medical Research. 22 (1), 42 (2017).
- Boué, S. M., et al. Evaluation of the estrogenic effects of legume extracts containing phytoestrogens. Journal of Agricultural and Food Chemistry. 51 (8), 2193-2199 (2003).
- Klinge, C. M. Estrogen receptor interaction with estrogen response elements. Nucleic Acids Research. 29 (14), 2905-2919 (2001).
- Nishikawa, J. -. i., et al. New screening methods for chemicals with hormonal activities using interaction of nuclear hormone receptor with coactivator. Toxicology and Applied Pharmacology. 154 (1), 76-83 (1999).
- . . Human Estrogen Receptor Beta (ERb; ESR2; NR3A2) Reporter Assay System. , (2020).
- Wasserman, M. D., et al. Estrogenic plant foods of red colobus monkeys and mountain gorillas in uganda. American Journal of Physical Anthropology. 148 (1), 88-97 (2012).
- Vivar, O. I., Saunier, E. F., Leitman, D. C., Firestone, G. L., Bjeldanes, L. F. Selective activation of estrogen receptor-β target genes by 3, 3′-diindolylmethane. Endocrinology. 151 (4), 1662-1667 (2010).
- Whitten, P. L., Patisaul, H. B. Cross-species and interassay comparisons of phytoestrogen action. Environmental Health Perspectives. 109, 5-20 (2001).
- Di Gioia, F., Petropoulos, S. A. . Advances in Food and Nutrition Research. , (2019).
- Lutz, I., Kloas, W. Amphibians as a model to study endocrine disruptors: I. Environmental pollution and estrogen receptor binding. Science of The Total Environment. 225 (1), 49-57 (1999).
- Felcyn, J. R., Davis, J. C. C., Tran, L. H., Berude, J. C., Latch, D. E. Aquatic Photochemistry of Isoflavone Phytoestrogens: Degradation Kinetics and Pathways. Environmental Science & Technology. 46 (12), 6698-6704 (2012).
- Jeng, Y. -. J., Kochukov, M. Y., Watson, C. S. Membrane estrogen receptor-alpha-mediated nongenomic actions of phytoestrogens in GH3/B6/F10 pituitary tumor cells. Journal of Molecular Signaling. 4, 2-2 (2009).
- Dixon, R. A. Phytoestrogens. Annual Review of Plant Biology. 55, (2004).
- Kuiper, G. G. J. M., et al. Interaction of Estrogenic Chemicals and Phytoestrogens with Estrogen Receptor β. Endocrinology. 139 (10), 4252-4263 (1998).
- Wasserman, M. D. . Feeding on Phytoestrogens: Implications of Estrogenic Plants for Primate Ecology. , (2011).
- Jefferson, W. N., Patisaul, H. B., Williams, C. J. Reproductive consequences of developmental phytoestrogen exposure. Reproduction. 143 (3), 247-260 (2012).
