Simple assay for screening phytoestrogenic compounds using the oestrogen receptor immobilised magnetite nanoparticles

Abstract

With increasing interests of phytoestrogens for their potential applications, a rapid and simple tool for screening these phytochemicals is still required. In this study, a simple assay to detect phytoestrogens was developed based on the competition binding between the tested samples and the fluorescently labelled oestrogen (E2) to the human ligand binding domain of oestrogen receptor (LBD‐ER) that was immobilised on the magnetite nanoparticles (MNPs). The 40‐kDa LBD‐ER peptide was produced in an Escherichia coli system. The synthesised 68.7‐nm MNPs were silanised and subsequently covalently linked to the C‐terminus of LBD‐ER peptide. The LBD‐ER immobilised MNPs demonstrated the specific binding for the standard E2 with the equilibrium dissociation constant of 9.56 nM and the binding capacity of 0.08 pmol/1 mg of the MNPs. The LBD‐ER immobilised MNPs could evaluate oestrogenic activity of the extracts of Asparagus racemosus and Curcuma comosa, the reported phytoestrogenic plants, but not progesterone (P4) and Raphanus sativus extract, the negative controls. The results of this work clearly demonstrated a potential assay for detecting phytoestrogens of crude plant extracts, which is simple and easily adapted to a high throughput format.

1 Introduction

Increasing numbers of synthetic and natural compounds that exhibit the activity interfering the normal functions of the human endocrine system have received many interests regarding their effects and applications to human health. Positive and negative effects of these compounds rely on their ability to compete with endogenous oestrogens for binding the receptors and subsequently cause a promotion or inhibition of oestrogenic responses [1]. These compounds are known as environmental endocrine disrupting chemicals (EDCs). Among these EDCs, natural phytochemicals and their active metabolites exhibiting a similar activity to a mammalian oestrogen are called phytoestrogens. Phytoestrogens are present in numerous dietary supplements and widely marketed as natural products for an alternative oestrogen replacement therapy [2, 3, 4, 5]. With increasing interests of their applications, plants containing high content or new phytoestrogens are investigated. Various in vivo (whole animal‐based) and in vitro (non‐whole animal based) assays are developed to evaluate the oestrogenic activity of phytochemicals. Examples are an uterotrophic assay [6], an evaluation of oestrogenic effects in female organs [7], a ligand‐oestrogen receptor (ER) binding assay [8], an ER‐promoter binding assay [9], an ER‐co‐activator binding assay [10], a transactivation assay [11], a gene expression analysis [12], a biochemical assay [13] and an E‐screen assay [14]. In vivo assays have advantages for determining complex responses of the whole body to the tested phytoestrogens, which are necessary for improving and facilitating extrapolation from in vivo animal data to the human situation [7]. However, in vivo assays are generally expensive, intensive, time‐consuming and require euthanasia of the animals [15, 16]. For in vitro assays, the principle of the assays is based on the specific interaction of the tested phytochemicals and ERs [7]. These methods are often used for a screening purpose because of a feasibility to handle numerous samples with a short analysis time. Nevertheless, an in vitro method with the higher sensitivity and shorter analysis time is still a challenge.

Magnetite nanoparticles (MNPs) have a great potential in biomedical applications, such as site‐specific drug delivery [17], hyperthermia for cancer therapy [18], magnetic resonance imaging and bio‐separation [19]. Magnetic separation of biomolecules is one of the important applications of MNPs due to a feasibility to conjugate a bait molecule on MNPs for a simple and rapid separation of target molecules by using an external magnetic separator [20]. This work, therefore, is interesting in developing a simple assay to evaluate the oestrogenic activity of phytochemicals by using MNPs and a ligand binding domain (LBD) of ER alpha. The principle of the assay to detect oestrogenicity of crude phytochemicals is based on a competitive binding between the fluorescently labelled standard oestradiol (E2) and the tested phytochemicals to the LBD‐ER immobilised MNPs. In addition to a high sensitivity, this developed assay can analyse many samples with reducing time due to a magnetic property of MNPs that allows a fast separation of the bound phytochemicals upon a use of an external magnet. Furthermore, this assay can evaluate the oestrogenicity of each phytochemical by directly comparing its binding ability for ER with that of the standard E2.

2 Materials and methods

2.1 Chemicals

Iron (III) chloride hexahydrate (FeCl₃ ·6H₂ O), tetraethyl orthosilicate (TEOS), N‐(2‐amino‐ethyl)‐3‐aminopropyl‐trimethoxysilane (APTES), 1‐ethyl‐(3‐3‐dimethylaminopropryl) carbodiimide hydrochloride (EDC) and ammonium hydroxide were purchased from Merck (Germany). Iron (II) sulphate heptahydrate (FeSO₂ ·7H₂ O), 17β‐oestradiol (E2), β‐oestradiol 6‐(O‐carboxymethyl)oxime: bovine serum albumin (BSA) fluorescein isothiocyanate conjugate (E2‐BSA‐FITC), N‐hydroxysuccinimide (NHS) were obtained from Sigma (USA). Progesterone (P4) was obtained from Tokyo Chemical Industry (Japan). All chemicals used were of analytical grade.

2.2 Production of LBD‐ER protein

DNA fragments encoding LBD‐ER (alpha type) were amplified by reverse transcription‐polymerase chain reaction (RT‐PCR) using a total RNA of MCF‐7 cells (human breast adenocarcinoma cell line). To synthesise cDNA, the reverse transcription of 1 μg of total RNA was performed using oligo(dT) priming and SuperScript^® III reverse transcriptase (Invitrogen, USA) in a total reaction of 20 μl according to the manufacturer's instructions. The cDNA was used as the template to amplify the DNA fragments encoding 344 amino acid residues of LBD‐ER (amino acids 252–595, Accession number NM_000125.3) by PCR. The PCR reaction (50 μl) contained 1 μl cDNA, 250 μM of dNTPs, 1× PCR buffer, 2.5 units of Taq DNA polymerase (RBC, Taiwan) and 1 μM each primer (5′–AAAGGTGGGATACGAAAAGAC–3′ and 5′–TCAGACTGTGGCAGGGAAAC–3′). The PCR condition contained one cycle at 94°C (2), followed by 30 cycles at 94°C (30 s), 60°C (30 s) and 72°C (90 s). The amplified DNA was ligated into the expression vector pQE‐30UA (Qiagen, USA) and subsequently transformed into Escherichia coli strain M15 [pREP4] (Qiagen, USA). After the positive clone was selected, the recombinant plasmid DNA was sequenced by the First Based DNA Sequencing Services (First Based Laboratories Sdn Bhd, Malaysia) to confirm its open reading frame and correct sequence.

To obtain the recombinant LBD‐ER peptide, a single bacterial colony was cultured in Luria–Bertani broth containing ampicillin (100 μg/ml) and kanamycin (100 μg/ml) at 37°C in an incubator shaker at 200 rpm overnight as a starter. The bacterial culture was subsequently diluted in a fresh medium (1:10 v/v) and grew until an OD₆₀₀ reached 0.6–0.8. Isopropyl‐β‐D‐1‐thiogalactopyranoside (IPTG, 1 mM) was added and the culture was further incubated for 6 h at 30°C. Cells were harvested by centrifugation at 6000 × g for 10 min. The cell pellet was resuspended in a binding buffer (20 mM sodium phosphate, 500 mM sodium chloride, 20 mM imidazole, pH 7.4 and 0.2 mg/ml lysozyme) and incubated at 37°C for 1 h. After lysing cells by sonication, the cell lysate was centrifuged at 10,000 × g for 10 min at 4°C and the supernatant was applied to a His‐GraviTrap column (GE Healthcare, USA). After the peptide was eluted with an elution buffer (20 mM sodium phosphate, 500 mM sodium chloride, 20 mM imidazole, pH 7.4), it was dialysed against 1× phosphate buffered saline (PBS), pH 7.4. The concentration of LBD‐ER peptide was determined using Quick Start^TM Bradford protein solution (Biorad, USA). The peptide was visualised by a 12.5% sodium dodecyl sulphate‐polyacrylamide gel electrophoresis (SDS‐PAGE) and stained with Coomassie brilliant blue G‐250.

2.3 Synthesis and surface modification of MNPs

MNPs were synthesised via a co‐precipitation method [21]. Ferrous sulphate (6.46 g) and ferric chloride (9.02 g) were prepared in 200 ml of 1% polyethylene glycol 4000 to make a final concentration of 15 mM of each iron salt. The solution was incubated at 80°C under vigorous stirring for 30 min. Then ammonium hydroxide (0.3 M) was dropped slowly into a mixture until the pH was titrated to 11.0 and further incubated at 80°C for 90 min under a protection of nitrogen gas. After the reaction mixture was cooled to room temperature, MNPs were isolated by centrifugation at 10,000×g for 10 min and washed with deionised water until the pH reached 7.0.

To immobilise LBD‐ER on MNPs, the protocol was prepared by a covalent immobilisation reaction according to Zhu et al. [22]. MNPs (5 g) were dispersed in water (16 ml) and subsequently mixed with ammonium hydroxide (2 ml) and ethanol (80 ml). TEOS (0.8 ml) was added dropwise to the solution under vigorous stirring at room temperature for 12 h. The silanised MNPs were collected by centrifugation at 10,000×g for 10 min and washed several times with deionised water. They were then dispersed in 54 ml ethanol before 1.62 ml APTES (3%) was slowly added. The reaction was incubated at room temperature for 2 h under nitrogen gas. After washing three times with ethanol, the modified MNPs were dispersed in 2.5% glutaraldehyde in 1× PBS under stirring for 5 h. Before the LBD‐ER immobilisation, carboxyl groups in silanised MNPs were activated using EDC/NHS. In brief, 200 mM EDC and 50 mM NHS were mixed with 3 ml of 1× PBS containing silanised MNPs (20 mg). The activated MNPs were washed with 1× PBS several times prior to incubating with various concentrations of LBD‐ER peptide at 4°C for 12 h to obtain an optimised concentration of the peptide. The LBD‐ER immobilised MNPs were magnetically separated and washed several times with 1× PBS until no free LBD‐ER was detected in the washing solution. The final LBD‐ER immobilised MNPs were redispersed in 1× PBS and stored at 4°C. The amount of immobilised LBD‐ER peptide was evaluated by using Quick Start^TM Bradford protein solution. In this work, BSA protein was also immobilised on MNPs to produce the BSA immobilised MNPs as the negative control.

2.4 Characterisation of the LBD‐ER immobilised MNPs

Morphology of MNPs and LBD‐ER immobilised MNPs were observed using a JEOL JSM 7800F field‐emission scanning electron microscope (FE‐SEM; JEOL, USA) provided with Schottky type field emission and lower electron detector (LED) at an accelerating voltage of 10 kV. All samples were dried, ground to a fine powder and sputter‐coated with gold immediately before observation. The average diameter of MNPs was determined from the SEM images at random locations (n  = 1000) using an ImageJ analysis (NIH, USA). The number‐average diameter (D _n), weight‐average diameter (D _w) and polydispersity index (PDI) were calculated from the following equations, where D_i is the diameter of the measured particles and N_i is the number of particles measured [23].

$$D_{\mathrm{n}}=\mathrm{\Sigma }{n}_i{D}_i/n_i$$ (1)

$$D_{\mathrm{w}}=\mathrm{\Sigma }{n}_i{D}_i^4/\mathrm{\Sigma }{n}_i{D}_i^3$$ (2)

$$\mathrm{PDI}=D_{\mathrm{w}}/D_{\mathrm{n}}$$ (3)

Attenuated total reflectance Fourier transform infrared spectroscopy (ATR‐FTIR) spectrophotometer (Bruker, USA) was used to identify functional groups of original MNPs, LBD‐ER peptide and LBD‐ER immobilised MNPs over a range of 400 to 4000 cm⁻¹ at a resolution of 4‐cm⁻¹ with 64 co‐added scans. Blackman–Harris three‐term apodisation, power‐spectrum phase correction and zero‐filling factor of 2 were set as default acquisition parameters using an OPUS 7.2 software suite (Bruker, USA). Background spectra of a clean ATR surface were obtained prior to each sample measurement using the same acquisition parameters.

A vibrating sample magnetometer (VSM; Lake Shore, USA) was used to measure the magnetic properties of MNPs and LBD‐ER immobilised MNPs. Samples were put down properly within sensing coils before they were measured at room temperature with a magnetic field in a range of −10,000 to 10,000 Oe, where parameters such as the saturation magnetisation (M _s), the remanent magnetisation (M _r) and the coercivity (H _c) were assessed.

2.5 Plant extraction

Roots of two phytoestrogenic plants (Asparagus racemosus and Curcuma comosa) and one non‐phytoestrogensic plant (Raphanus sativus) were used. Plant extraction method was modified from Martins et al. [24]. Air‐dried plant samples were ground to fine powder and 1 g each sample was extracted in 20 ml ethanol (50%) in a water‐bath at 60°C for 30 min. Extracts were filtered through a Whatman No. 1 paper and the filtrates were evaporated at 60°C. The extracted samples were freeze‐dried at −110°C and stored at −20°C until a further analysis.

2.6 Optimised conditions for a specific binding of the produced MNPs

The ability of the immobilised LBD‐ER binding to E2 was evaluated. LBD‐ER immobilised MNPs were incubated with the standard E2‐BSA‐FITC solution (0.5 × 10⁻¹⁰ to 1 × 10⁻⁵ M) for 60 min at 37°C. After that, they were magnetically separated and washed three times with 1× PBS. The bound E2‐BSA‐FITC on the MNPs was determined by a fluorescence microplate reader (Molecular devices, USA) using the excitation and emission wavelengths at 485 and 538 nm, respectively. A similar experiment was conducted using the BSA immobilised MNPs as the negative control.

To determine the optimal washing condition for a removal of non‐specific binding, the LBD‐ER immobilised MNPs (1 mg) was incubated with the standard E2‐BSA‐FITC (1.0 × 10⁻⁶ M) for 60 min at 37°C. The LBD‐ER immobilised MNPs were drawn by an external magnet and the remained solution was removed. The MNPs were washed with 1× PBS (200 µl) and the washed solution was collected. The washing step was performed for 12 times. The amount of washed E2‐BSA‐FITC in each washing step and the amount of bound E2‐BSA‐FITC were measured by a fluorescence microplate reader.

For a saturation binding assay, LBD‐ER immobilised MNPs (1 mg) in 1× PBS were incubated at 37°C for 30 min with E2‐BSA‐FITC (2–20 nM) in an absence or presence of a 500‐fold excess of non‐fluorescently labelled E2. The bound and free fluorescent ligands were measured after the LBD‐ER immobilised MNPs were drawn by an external magnet. The specific binding was evaluated by a subtraction of a non‐specific binding (bound fluorescent ligands in a condition with an excess of unlabelled E2) from a total binding. The specific binding data were subjected to the Hill equation and linear Scatchard transformation to determine the equilibrium dissociation constant (K _d) and binding capacity (B _max) using the Origin Pro 2015 (OriginLab Corp., USA) software program.

2.7 LBD‐ER immobilised MNPs to determine oestrogenic activity

The ability of LBD‐ER immobilised MNPs to determine oestrogenic activity of a tested sample based on a competitive binding was investigated. LBD‐ER immobilised MNPs (1 mg) were incubated with 1.0 × 10⁻⁶ M E2‐BSA‐FITC at 37°C for 1 h before the crude plant extracts (15.6–2000 µg/ml), the standard E2 (10⁻¹² –10⁻⁴ M) or the negative control progesterone (P4, 10⁻¹² –10⁻⁴ M) were added. In this work, three crude plant extracts were used, which were derived from two phytoestrogenic plants (A. racemosus and C. comosa) and one non‐phytoestrogensic plant (R. sativus). After incubating for 30 min, the LBD‐ER immobilised MNPs were drawn by an external magnet. The collected LBD‐ER immobilised MNPs were washed with 1 × PBS and the bound E2‐BSA‐FITC was measured. Fluorescent intensity indicated the oestrogenic activity of the tested samples, which competed with the standard E2‐BSA‐FITC for binding the LBD‐ER immobilised MNPs. The fluorescent data were fitted with a non‐linear model for determining a dose response and a concentration of 50% inhibition (IC₅₀). The IC₅₀ value is described as the concentration of a tested competitive sample to occupy 50% of the LBD‐ER, thus causing a 50% reduction of fluorescent intensity.

2.8 Determination of the specific binding of plant extracts to LBD‐ER immobilised MNPs

To determine the specific binding, the plant extracts of A. racemosus, C. Comosa and R. sativus (1 mg in 1 ml of 1× PBS containing 2% DMSO) were incubated with the LBD‐ER immobilised MNPs (1 mg) at 37°C. In addition, the standard E2 (1 mg) and P4 (1 mg) were used as the positive and negative controls, respectively. After incubating for 30 min, the LBD‐ER immobilised MNPs were drawn to one side of the tube by applying an external magnet. After the solution was removed, the LBD‐ER immobilised MNPs were washed with 1× PBS (1 ml) for three times to remove non‐specific binding compounds. The LBD‐ER immobilised MNPs were then boiled at 100°C for 5 min to denature the LBD‐ER protein, thus releasing the bound compounds. The content of bound compounds was determined by measuring the absorbance at 280 nm, commonly used to determine the relative contents of E2, P4 and phytoestrogens [25, 26].

2.9 Statistical analysis

The experiments were done in triplicate and results were expressed as mean ±SD. Statistical analyses were completed using SPSS version 18.0 (SPSS, USA) and the treatments were considered statistically significant when the p value was less than 0.05. Differences between the samples were analysed by one‐way analysis of variance (ANOVA) and the Tukey's honestly significant difference was also used for pairwise analysis.

3 Results and discussion

3.1 Production of LBD‐ER protein

The DNA encoding LBD‐ER protein was synthesised by RT‐PCR. The amplified DNA of 1035 bp was obtained (Fig. 1 a), which subsequently cloned into the protein expression vector and finally yielded the recombinant plasmid, pQE‐LBD‐ER. The insert DNA sequence of the recombinant plasmid was confirmed for its correct sequence and open reading frame (data not shown). The recombinant LBD‐ER protein was obtained by inducing the transformed E. coli containing pQE‐LBD‐ER with 1 mM IPTG and subsequently purifying the recombinant protein via a Ni⁺ ‐affinity column chromatography. The expressed proteins in E. coli cells analysed on a 12.5% SDS‐PAGE gel are shown in Fig. 1 b. The predominant band of a molecular mass of approximately 40 kDa was clearly shown in the bacterial lysate under an IPTG‐induction (lane 2), but not in the non‐IPTG‐induced one (lane 1). After a purification, a single 40‐kDa protein (lane 3) was obtained, which well corresponded to the predicted size (40.9 kDa) of the recombinant LBD‐ER protein. The sequence of LBD‐ER protein covers the 12 α‐helix region that can arrange into a tertiary pocket structure for an oestrogenic ligand binding [27]. Therefore, this protein was used as a specific bait for compounds exhibiting oestrogenic activity in this study.

Fig. 1.

Amplified LBD‐ER DNA and the purified LBD‐ER peptide

(a) Amplified LBD‐ER DNA of 1035 bp (arrow) on a 0.8% agarose gel. Lane M: 1 kb plus DNA ladder (Invitrogen, USA), (b) Expressed proteins in E. coli cells separated on a 12.5% SDS‐PAGE gel stained with Coomassie Brilliant Blue. Lane M: PageRuler™ unstained protein ladder (10–200 kDa). Lane 1: Bacterial cell lysate without an IPTG induction. Lane 2: Bacterial cell lysate with an IPTG induction. Lane 3: Purified recombinant LBD‐ER peptide of 40 kDa (arrow)

3.2 Preparation and characterisation of LBD‐ER immobilised MNPs

In this work, MNPs were synthesised via a co‐precipitation reaction and their morphology was observed by SEM. Fig. 2 a shows SEM micrographs of the obtained MNPs, which were spherical. The number‐average diameter (D _n), the weight‐average diameter (D _w) and the PDI of the MNPs were 68.7 nm, 79.8 nm and 1.160, respectively. The synthesised MNPs were subsequently silanised and activated to form a covalent link to a C‐terminus of LBD‐ER peptide. The optimised ratio was to use 250 μg of the peptide per 1 mg of the activated MNPs (data not shown). After immobilising the LBD‐ER peptide, the modified MNPs remained round but slightly increased in size (D _n  = 94.5 nm and D _w  = 110.0 nm) as seen in Fig. 2 b. The PDI value was 1.164. Based on the PDI values in a range of 1.1–1.2, the distribution of the produced MNPs were considered as the nearly monodisperse distribution [28].

Fig. 2.

SEM images and size distribution of MNPs and LBD‐ER MNPs

(a) Synthesised MNPs, (b) LBD‐ER immobilised MNPs

The LBD‐ER immobilisation on the surface of MNPs was confirmed by ATR‐FTIR analysis. The ATR‐FTIR spectrum of MNPs revealed a single, strong absorption peak at 543 cm⁻¹ (Fig. 3) assigned to the Fe–O bond vibration [29]. The characteristic absorption peaks of LBD‐ER peptide were observed at 1400, 1564, 1627 and 3430 cm⁻¹ attributed to the stretching vibration of C–C in an aromatic ring, amide II, amide I and N–H stretching of NH₂, respectively. For the LBD‐ER immobilised MNPs, several spectral peaks were observed in addition to the absorption peak of Fe–O at 549 cm⁻¹. The spectral peak at 1053 cm⁻¹ was attributed to the un‐symmetric and symmetric linear vibration of Si–O–Si bonding [30], indicating the existence of SiO₂ in the particles. The presence of LBD‐ER peptide was indicated from the characteristic peaks of C–C stretching at 1402 cm⁻¹, amide II at 1541 cm⁻¹, amide I at 1631 cm⁻¹ and N–H stretching of NH₂ at 3430 cm⁻¹ [30, 31]. In addition, the characteristic absorption band for C–N stretching at 1213 cm⁻¹ [32], suggesting the peptide bond between each amino acid unit as well as the peptide bond between silanised MNPs and the LBD‐ER peptide.

Fig. 3.

FTIR spectra of LBD‐ER, MNPs and LBD‐ER MNPs in a range of 400–4000 cm⁻¹

Magnetic behaviour of the MNPs before and after immobilisation with the peptide was analysed via a hysteresis curve as shown in Fig. 4. Both MNPs and LBD‐ER immobilised MNPs reached the saturation at the field of 9.7 kOe. The saturation magnetisation (M _s), the remanent magnetisation (M _r) and the coercivity (H _c) for the MNPs are 50.2 emu·g⁻¹, 1.0 emu·g⁻¹ and 18.0 Oe, respectively, and those for the LBD‐ER immobilised MNPs are 40.4 emu·g⁻¹, 0.8 emu·g⁻¹ and 17.1 Oe, respectively. Both MNPs showed ferromagnetic behaviour as indicated by the saturated state, low remanence and low coercivity on the magnetisation curve [33]. The lower values of all magnetic parameters of the LBD‐ER immobilised MNPs are likely due to the coated layers of protein on the MNPs [31, 34]. Nevertheless, the magnetic properties of the LBD‐ER immobilised MNPs are important for an application in this work, which is to facilitate a separation of ligand‐bound MNPs by applying an external magnet.

Fig. 4.

Hysteresis loops of MNPs and LBD‐ERimmobilised MNPs

3.3 Optimised conditions for a specific binding of the modified MNPs

The ability of the immobilised LBD‐ER peptide to bind E2‐BSA‐FITC is shown in Fig. 5. The immobilised LBD‐ER peptide demonstrated a dose‐dependent binding to E2‐BSA‐FITC, clearly seen in a range of 10⁻⁸ –10⁻⁶ M. No significant increase of binding was observed at the greater concentrations than 10⁻⁶ M of E2‐BSA‐FITC. In this experiment, the BSA immobilised MNPs was used as the negative control protein and no binding to E2‐BSA‐FITC was detected. It was noted that the E2‐BSA‐FITC might not be as efficient as E2 for binding to LBD‐ER due to a bulky structure of attached BSA [35]. However, it could specifically bind to LBD‐ER and frequently used as a tracer in a competitive binding assay [36].

Fig. 5.

Binding of the LBD‐ER immobilised MNPs to E2‐BSA‐FITC compared with the BSA immobilised MNPs

The optimal washing condition for removing non‐specific binding was determined and the results are shown in Fig. 6. The amounts of E2‐BSA‐FITC that remained binding to the LBD‐ER immobilised MNPs or the BSA immobilised MNPs after several washes were evaluated by fluorescence intensity. Non‐specific binding of E2‐BSA‐FITC on the LBD‐ER immobilised MNPs was removed by washing repeatedly and the amount of bound E2‐BSA‐FITC became constant after washing twice (Fig. 6 a). The amount of washed E2‐BSA‐FITC was also measured as seen in Fig. 6 b. Well corresponded to the results in Fig. 6 a, the fluorescence intensity of the washed E2‐BSA‐FITC in the supernatant was barely detected after two times of washing. Similar experiments were performed using the negative control, the BSA immobilised MNPs. It was clearly shown that the BSA immobilised MNPs rarely bound to E2‐BSA‐FITC and a twice‐washing step was sufficient to remove any non‐specific binding.

Fig. 6.

Effect of washing times on non‐specific binding between the LBD‐ER immobilised MNPs and E2‐BSA‐FITC as compared with the negative control, BSA immobilised MNPs

(a) Fluorescence intensity of bound E2 on the MNPs, (b) Fluorescence intensity of washed E2 in the supernatant

3.4 Binding affinity of the LBD‐ER immobilised MNPs to E2

The binding affinity of the LBD‐ER immobilised MNPs was evaluated from a saturation curve of its ligand, E2‐BSA‐FITC (Fig. 7), which the equilibrium dissociation constant (K _d) was determined by a Scatchard analysis. The calculated K _d value was 9.56 × 10⁻⁹ M and the binding capacity (B _max) value was 0.08 pmol/1 mg of the LBD‐ER immobilised MNPs. This K _d value was well corresponded to the previous reports in a range of 0.1 × 10⁻⁹ M to 25 × 10⁻⁹ M [37, 38, 39, 40].

Fig. 7.

Saturation binding of E2‐BSA‐FITC to the LBD‐ER immobilised MNPs. The specific binding was evaluated by a subtraction of a non‐specific binding (bound fluorescent ligands in a condition with an excess of unlabelled E2) from a total binding. Inset is the Scatchard analysis of the binding data. The binding capacity, B_max  = 0.08 pmol/1 mg LBD‐ER immobilised MNPs and the equilibrium dissociation constant, K_d  = 9.56 × 10⁻⁹ M, were calculated. Each point represents the mean ± standard deviation in the saturation binding graph from three independent assays

3.5 LBD‐ER immobilised MNPs for determining oestrogenic activity

The LBD‐ER immobilised MNPs were used to evaluate the oestrogenic activity of the standard E2, which the assay was based on a competitive binding between the tested standard E2 and the bound E2‐BSA‐FITC (tracer) for the LBD‐ER immobilised MNPs. In this experiment, progesterone (P4), the steroid hormone with a similar structure to E2, was used as the negative control. The results are shown in Fig. 8 a. The standard E2 exhibited a specific binding to the LBD‐ER immobilised MNPs by competing with the bound E2‐BSA‐FITC and demonstrated a sigmoidal curve with the IC₅₀ value at 1.26 × 10⁻⁹ M. In contrast, P4 showed no ability to compete with the E2‐BSA‐FITC for binding the LBD‐ER immobilised MNPs as seen in Fig. 8 b. The developed assay demonstrated a specific detection of E2 in a nanomolar concentration, which was similar to the radiolabelled displacement competitive inhibition assays (8.99 × 10⁻¹⁰ to 2.3 × 10⁻⁹ M) [39, 41, 42, 43], but greater than the reporter gene assays (∼10⁻⁷ M) [44], E‐screen assays (∼10⁻⁶ M) and yeast oestrogen screen assays (∼10⁻⁵ M) [45]. As compared with above assays, the developed assay was easily and rapidly performed since MNPs could be trapped on one side of the tube by applying an external magnet, thus making fast separating and washing processes.

Fig. 8.

Competitive binding of the tested compounds to the bound E2‐BSA‐FITC on the LBD‐ER immobilised MNPs

(a) E2 and P4, (b) Three crude plant extracts: R. sativus, A. racemosus and C. comosa. The inset is the competitive binding curve in a range of 500–2000 µg/ml of plant extracts

Subsequently, the LBD‐ER immobilised MNPs were used to determine the oestrogenic activity of crude extracts derived from two phytoestrogenic plants, A. racemosus and C. comosa [46, 47] and one non‐phytoestrogenic plant, R. sativus. Crude extracts of A. racemosus and C. comosa demonstrated a competitive binding with the E2‐BSA‐FITC for the LBD‐ER immobilised MNPs as indicated by a greater reduction of detecting fluorescence intensity when higher concentrations of the plant extracts were included (Fig. 8 b). The IC₅₀ values of the crude plant extracts were 1330.4 ± 30.3 and 1293.3 ± 22.4 µg/ml, respectively (Table 1). The competitive binding of the A. racemosus and C. comosa extracts also demonstrated a sigmoid binding curve, as clearly seen in a range of 500–2000 µg/ml (inset of Fig. 8 b), suggesting a single, affinity binding site of the tested phytochemicals for the ER in accordance with their oestrogenic activity [48]. In contrast, no oestrogenic activity of R. sativus, the negative control, was detected. Although oestrogenic activity of A. racemosus and C. comosa was not previously reported by any in vitro assay, their oestrogenic activities were reported by in vivo uterotrophic assay. Both extracts demonstrated oestrogenic effects in genital organs and in mammary glands in rats but at least 300 mg of A. racemosus extract and 125 mg of C. comosa extract per kilogram of a bodied weight of rat were required for this detection [49, 50].

Table 1.

IC₅₀ values of tested compounds using the LBD‐ER immobilised MNPs and an E2‐BSA‐FITC tracer

Compound	IC50 a
E2	1.26 × 10−9 (M)
P4	nbb
R. sativus extract	nbb
A. racemosus extract	1330.4 ± 30.3 µg/mlc
C. comosa extract	1293.3 ± 22.4 µg/mlc

^a Concentration for 50% inhibition of E2‐BSA‐FITC.

^b No binding under conditions employed.

^c Denotes p  < 0.05 using the statistical test ANOVA. Data represent the mean and standard deviation of n  = 5.

It was important to point that it could not rule out the possibility of some unknown components in the plant extracts could inhibit the E2‐BSA‐FITC binding, thus partially causing the reduction of fluorescent signal in Fig. 8 b. However, it was more likely that the reduction of fluorescent signal was mainly caused by the competitive binding of phytochemicals derived from A. racemosus and C. comosa extracts. To confirm this point, the specific binding of LBD‐ER immobilised MNPs to both plant extracts were tested. After LBD‐ER immobilised MNPs were allowed to bind each plant extracts (1 mg/ml) for 30 min, they were washed three times to remove non‐specific binding. The bound phytochemicals of each plant extracts were evaluated by the absorbance at 280 nm. Fig. 9 shows the bound phytochemicals and the bound standard E2 that were released from the LBD‐ER immobilised MNPs, suggesting that the A. racemosus and C. comosa extracts contained the phytochemicals that specifically bound to the LBD‐ER. Thus, these phytochemicals could compete with the standard E2 for binding the LBD‐ER immobilised MNPs. In contrast, bound phytochemicals of R. sativus extract and the negative control P4 were barely detected (A280 at 0.006 and 0.007, respectively), suggesting no or very low binding of R. sativus extract and P4 for the LBD‐ER immobilised MNPs.

Fig. 9.

Relative contents of released E2, P4, A. racemosus (AR) compounds, C. comosa (CC) compounds and R. sativus (RS) compounds that were specifically bound the LBD‐ER immobilised MNPs

4 Conclusion

A rapid and high throughput assay to detect oestrogenic activity based on a competitive binding to LBD‐ER immobilised MNPs was developed as a tool for an initial screening of oestrogenic compounds, including phytoestrogens of crude plant extracts. The binding affinity of the LBD‐ER immobilised MNPs to the standard E2 was in a range of nanomolar and the binding capacity was as low as 0.08 pmol of E2 per 1 mg MNPs. With the magnetic property, the LBD‐ER immobilised MNPs also allowed a simple and fast separation of unreacted chemicals through multiple washing steps by applying an external magnet, thus decreasing a non‐specific binding to the LBD‐ER immobilised MNPs. In addition, the developed assay demonstrated a specific detection of oestrogenic activity of crude plant extracts.