Fluorescence immunoassays using CdSe/ZnS core/shell quantum dots for the determination of progesterone in human serum

In this study, two heterogeneous fluorescence immunoassays using CdSe/ZnS quantum dot (QD) to label anti-progesterone antibody (P4Ab) for the determination of progesterone (P4) were performed in the wells of a 96-well microtiter plate. First, P4Ab was conjugated to hydrophilic CdSe/ZnS QDs via ethyl-3-(dimethylaminopropyl) carbodiimide(EDC)-Nhydroxysuccinimide) chemistry(NHS) (QDs-P4Ab conjugates). The QDs-P4Ab conjugate was employed as a second antibody in a sandwich assay, where the P4Ab was immobilized onto the 3-aminopropyltrimethoxysilane (APTMS) sol-gel membrane of the wells of a 96-well microtiter plate, and P4 was bound between the immobilized P4Ab and the QDs-P4Ab conjugate. In this assay, the fluorescence intensity of the QDs increased with increasing P4 concentrations. This assay had a detection limit of 553.9 pg/ml and a sensitivity of 18,251.96 pg/ml with a linear range of 2,184.6 – 117,082 pg/ml. In the direct binding assay, P4 was directly bound to the QDs-P4Ab conjugates immobilized onto the APTMS sol-gel membrane of the wells of a 96-well microtiter plate. In this direct binding assay the fluorescence intensity of the QDs decreased with increasing P4 concentrations, and this assay had a linear range of 28.95 – 26,607.7 pg/ml with a detection limit of 3.32 pg/ml and a sensitivity of 987.24 pg/ml. These fluorescence immunoassays have been successfully applied for the determination of P4 in real human serum, and the results were well correlated with those of a certified radioimmunoassay (RIA) method.


Introduction
Immunoassay is perhaps the methodology most frequently used to measure biological compounds in translational and clinical research [1]. The methods and categories of immunoassays have been reviewed by scientists such as Ashihara et al. [1], D.S Hage [2], and I.A. Darwish [3]. In these methodologies, a number of new techniques have been developed to quantify biological substrates that differ from the primary detection method by having high sensitivity and expanding the detection limit. One of these is fluorescence immunoassays which utilize fluorescent compounds as immuno-labels [4][5][6].
Human serum measurement methods, such as the direct detection and monitoring of individual labeled antibodies and their reactions with antigens, are complicated by the fact that, in addition to organic ions and hydrophilic organic substances, about 100 different proteins are present in human serum. As a result, human serum samples emit strongly when they are irradiated in blue or green regions of the spectrum [6]. This prevents the detection of single chromophores, such as fluorescein or classic rhodamine dye, in this environment. Therefore, in the analysis of bio-samples, the use of fluorescent materials with emission wavelengths out of this spectral range is preferred. The use of fluorescent materials in immunoassays creates bovine serum samples were determined based on the direct attachment of P4Ab on a modified gold disk electrode coated with gold nanoparticles [17]. P4-coating antigen was immobilized on a glassy carbon electrode coated with thionine-graphene oxide composites in the development of a homogeneous immunoassay [18]. In the development of a homogeneous QDbased fluorescence immunoassay, QD-antigen conjugates can be employed with an unlabeled anti-progesterone antibody. Progesterone has been detected in milk by using a competitive immunoassay using QDs as fluorescent labels [19]. However, the use of QD-conjugated antigens for competition with free antigens means that the routine determination of P4 concentrations can become expensive.
From this point of view, CdSe/ZnS core/shell QDs have been employed as the label of anti-progesterone antibody (P4Ab) in immunoassays for P4 detection. Hydrophilic CdSe/ZnS QDs were conjugated with the anti-progesterone antibody (P4Ab) via the ethyl-3-(dimethylaminopropyl) carbodiimide (EDC)/N-hydroxysuccinimide (NHS) coupling method. The QDs-conjugated P4Ab (QDs-P4Ab conjugate) was used to develop two heterogeneous fluorescence immunoassays for the determination of progesterone in human serum. One heterogeneous assay was based on the direct binding of the sample (P4) to the QDs-P4Ab conjugate immobilized onto the 3-aminopropyltrimethoxysilane (APTMS) sol-gel membrane of the wells in a 96-well microtiter plate, while in our previous paper the QDs-P4Ab conjugate was immobilized onto the GA (a mixture of GPTMS (3-glycidoxypropyltrimethoxysilane) and APTMS) sol-gel membrane of the wells [20]. In the sandwich assay, the P4Ab was immobilized onto the APTMS sol-gel membrane of the wells and the QDs-P4Ab conjugates were employed as a secondary antibody after the progesterone was added. In this study these two heterogeneous immunoassays using the QDs-P4Ab conjugates have been systematically characterized and applied for the determination of progesterone in human serum, and they have been compared with the commercially available progesterone assay method of RIA.

Synthesis of hydrophilic CdSe/ZnS QDs
The synthesis of the CdSe/ZnS QDs was based on previously described theories with slight modification. CdSe nanoparticles were synthesized using the modified methods of Qu and Peng [21], while CdSe/ZnS QDs were synthesized following the method of Gaunt et al. [22]. For the hydrophilic surfactant coated CdSe/ZnS QDs, MPA was capped onto CdSe/ZnS QDs by following the protocols described in the literature with minimal modification [23][24]. The absorption or emission spectra of CdSe and CdSe/ZnS QDs were respectively determined by UV-Vis spectroscopy (Multiskan Spectrum, Thermo electron Co., Finland) and a fluorescence spectrophotometer (Model: F-4500, Hitachi Co., Japan).

Preparation and purification of the QDs-P4Ab conjugates
First, 200 μl solution of the MPA-capped QDs, 20 μl of EDC (30 mg/ml), and 20 μl of NHS (32 mg/ml) were mixed with a fixed amount of P4Ab in 10 mM phosphate buffer solution (PBS) in a glass tube. The coupling reaction was performed in the glass tube at 25 ℃ for 1 hr with gentle agitation. The QDs-P4Ab conjugates were confirmed by gel electrophoresis (2% agarose). A voltage of 100 V was applied along the gel for 30 min, and the gels were then photographed under UV-light. In this study, a fixed amount of CdSe/ZnS QDs (3.2 mg in 1 ml of 10 mM PBS) was conjugated to different concentrations (0, 62.5, 125, 250, 500, and 1000 ng/ml) of P4Ab. In addition, a fixed amount of P4Ab (0.25 mg in 1 ml of 10 mM PBS) was conjugated to different amounts (0, 0.4, 0.8, 1.6, 2.4, 3.2 mg/ml) of QDs.
The QDs-P4Ab conjugates were separated from unconjugated QDs or P4Ab with a Fast Protein Liquid Chromatography (FPLC) system (ÄKTA purifier, Amersham Pharmacia Biotech Co., USA) having a Sephadex G-100 column (30 cm long and 0.9 cm diameter). The column was first pre-equilibrated with PBS (0.05 M, pH 7.4) at 0.3 ml/min. The sample, which had a total volume of 500 μl was then loaded onto the column. A 280 nm UV-light source was used to check the flow from the column, and the eluted solution was collected for further analysis. The fluorescence spectra of the purified QDs-P4Ab conjugates were also determined using a fluorescence spectrophotometer (Model: F-4500, Hitachi Co., Japan).

Preparation of the APTMS sol-gel membrane and immobilization of the P4Ab or QDs-P4Ab conjugates
The APTMS sol-gel membrane in the well of a 96-well microtiter plate was prepared in the following manners [25]: 312.5 μl of APTMS was mixed in 687.5 μl of 99% ethanol solvent and polymerized by adding 40 μl of 35% hydrochloric acid. The APTMS sol-gel matrix was kept at room temperature for at least 3 hours under a shaking condition (180 rpm) prior to use. Then, 5 μl of the APTMS sol-gel solution was spread on the bottom of one well in a 96well microtiter plate and dried at room temperature for 30 min.
To immobilize the P4Ab (or QDs-P4Ab conjugates) onto a well in a 96-well microtiter plate, 5 μL of the APTMS sol-gel solution was first spread onto the bottom of a well and dried at room temperature for 30 min. Next, 10 μl of glutaraldehyde solution (1 wt%) and 100 μl of the P4Ab (or QDs-P4Ab conjugates) in 10 mM PBS (pH 7.4) were added to the well and incubated at 4 °C for 30 hours. After incubation, the well was washed with 10 mM PBS (pH 7.4) three times. Then, 100 μl of bovine serum albumin (BSA) solution (5 mg/ml) was added to the wells to block the unoccupied sites on the well surface, incubated for 3 hours, and washed with 10 mM PBS (pH 7.4) three times. The surface morphology of the APTMS sol-gel membrane containing the QDs-P4Ab conjugates was identified by an atomic force microscopy (AFM, Multimode IV, Veeco Instruments Inc., USA) and a scanning electron microscope (SEM, S-4700, Hitachi, Japan).

Sandwich assay
In the sandwich assay, 100 μl of progesterone (P4) dissolved in 10 mM PBS (pH 7.4) was added to a well that was immobilized with P4Ab (12.5 ng/100 μl/well) and incubated at 4 °C for 24 hours. After washing the well with 10 mM PBS (pH 7.4) at least three times, 100 μl of the P4Ab-QDs conjugates was added to the well and incubated for 24 hours. The fluorescence intensity of the well was measured with the microtiter plate reader (Saphire 2 , TECAN Co., Austria) at excitation/emission wavelength of 475 nm and 590 nm, respectively, with various concentrations of P4 solution (0-20,000 pg/ml), after washing the wells at least three times. The measurements of P4 were performed in triplicate. To dissociate the binding among P4, the immobilized P4Ab, and the P4Ab-QDs conjugates, 0.1 M glycine•HCl at pH 2.5-3.0 was employed as an elution buffer.
The effects on the assay performance of different amounts of the P4Ab immobilized onto a well surface via the APTMS sol-gel membrane were investigated with 6.25 ng and 12.5 ng in 100 μl per well in a 96-well microtiter plate.
The binding between P4 and the QDs-P4Ab conjugates was studied at different incubation times and with various P4 concentrations. First, the sample (P4) was incubated at 4 °C for 12 hours to bind to the immobilized P4Ab and the wells were washed with 10 mM PBS (pH 7.4). Then, the QDs-P4Ab conjugates were introduced into the wells and incubated for 3, 6, and 24 hours.
The interference of some components normally existing in real blood samples such as ascorbic acid, Na + , Ca ++ , K + , and different hormones (estrogen, testosterone) on the samples was examined in the P4 concentration range from 0 to 5,000 pg/ml. Ascorbic acid (0.1, 0.5, and 1.0 mM) and some ions (K + , Na + , and Ca 2+ in the concentrations of 50 mM, 100 mM, and 150 mM, respectively) in a sample were also investigated.
Three elution buffers were studied to separate the bonds among P4 and immobilized P4Ab and P4Ab-QDs conjugates. Specifically, 0.1 M glycine•HCl (pH 2.5-3.0) and 0.2 M NaOH (pH 13.1) were prepared and used as acidic and alkaline elution buffers, respectively. 50 mM Tris-HCl (pH 7.0) containing 2% sodium dodecyl sulfate (SDS) and 50 mM dithiothreitol (DTT) was also made and used for the dissociation. Each elution buffer was incubated in a well for either 5 or 10 minutes.
In this sandwich immunoassay a significant difference in the fluorescence intensities was also found in the P4 concentration range of 50 to 10,000 pg/ml. The fluorescence intensity for each sample was measured after an incubation time of 3 hours.

Direct binding assay
In this assay, 100 μl of progesterone (P4) dissolved in 10 mM PBS (pH 7.4) was added to each well immobilized with the QDs-P4Ab conjugates (e.g. conjugation of 0.32 mg CdSe/ZnS QDs added to 500 ng/ml P4Ab). After incubating the samples (P4) for 24 hours, the wells were washed with 10 mM PBS (pH 7.4) at least three times. The 96-well microtiter plate was then inserted into the measurement chamber of the microplate reader (Saphire 2 , TECAN Co., Austria), and the fluorescence intensity of each well was measured at excitation/emission wavelengths of 475 nm and 590 nm, respectively, in the P4 concentration range from 0 to 20,000 pg/ml. The measurement of P4 was performed in triplicate. The dissociation of the binding between P4 and the QDs-P4Ab conjugates was carried out with 0.1 M glycine•HCl at pH 2.5-3.0 as an elution buffer.
The long-term stability of the immobilized QDs-P4Ab conjugates was also evaluated using a well immobilized with 100 μl of the QDs-P4Ab conjugates. After incubating the samples (P4) at different P4 concentrations for 24 hours and washing each well with 10 mM PBS (pH 7.4) several times, the fluorescence intensity of the well was measured every week. When the microplate was not being used, it was kept in a refrigerator.
Optimum amounts of the QDs-P4Ab conjugates immobilized onto the wells in a 96-well microtiter plate were also investigated with 50 μl, 100 μl, and 150 μl of the QDs-P4Ab conjugates. After immobilizing 50 μl, 100 μl, and 150 μl of the QDs-P4Ab conjugates in wells, samples (P4) at different concentrations were introduced into the wells and incubated for 3 hours at room temperature. The fluorescence intensity of the wells was measured after washing each well with 10 mM PBS (pH 7.4) several times.
The binding between P4 and the QDs-P4Ab conjugates was also studied at different incubation times (3, 6, and 24 hours) and with various P4 concentrations.
The interference of some components such as ascorbic acid (0.1, 0.5, 1.0 mM), Na + (50, 100, 150 mM), estrogen, and testosterone (500 and 5000 pg/ml) on the binding between P4 and the immobilized P4Ab was investigated in the P4 concentration range from 0 to 5,000 pg/ml. The effects of pH on the progesterone measurement were also investigated.
The same three elution buffers as those used in the sandwich assay (0.1 M glycine•HCl (pH 2.5-3.0), 50 mM Tris-HCl (pH 7) containing 2% SDS and 50 mM DTT, and 0.2 M NaOH (pH 13.1)) were prepared and studied to dissociate the binding among the sample (P4) and the QDs-P4Ab conjugates. The elution buffer was also incubated in a well for either 5 or 10 minutes.
To remove some unbound components from a samle, such as other proteins, several washing buffers were prepared with the addition of surfactants such as Tween X20 (0.05 v/v%) and X100 (0.05 v/v% and 0.1 v/v%) to 10 mM PBS.
In this direct binding assay, a significant difference in the fluorescence intensities was found in the entire range of progesterone concentrations (50 -10,000 pg/ml). To ensure that the affinity reaction was completely finished, the fluorescence intensity for each sample was measured after an incubation time of 3 hours.

Radioimmunoassay (RIA)
Radioimmunoassay (RIA) for progesterone was performed using Progesterone 125 I RIA Kit (ICN, Biochemical Inc., Diagnostic Division, Costa Mesa, CA, USA) according to the manufacturer's protocol with minor modifications [26]. Radioactivity was determined using a gamma counter (Packard, Cobra II 5005, Meriden, CT, USA). All samples were run in triplicate and two sets of progesterone standards were included in each assay. The concentrations of progesterone were calculated with smart RIA (Packard, Meriden, CT, USA). The progesterone antiserum was developed in a rabbit using testosterone-17-β-hemisuccinate: BSA as an immunogene [27]. The lower limit of assay sensitivity for progesterone was 5 pg/ml.

Data analysis
Each set of data was obtained from at least three independent measurements and presented as mean ± standard deviation (SD). The difference in the sensitivity between the different assays or in the fluorescence intensity in the presence of different ions (K + , Na + ) was evaluated by one-way analysis of variance (ANOVA). The differences observed between samples were considered to be significant at p-values (probability) less than 0.05. The statistical data tests were conducting using the software InStat (vers.3.01, GraphPad Software Inc., San Diego, CA, USA). The normalized fluorescence intensity (Norm. Fl. intensity) represents the ratio of the fluorescence intensity enhancement, which is calculated by dividing the difference between the fluorescence intensity (FI) at a given progesterone concentration and the fluorescence intensity (FIo) at a concentration of 0.0 pg/ml by the fluorescence intensity (FIo) at a progesterone concentration of 0.0 pg/ml, i.e. (FI -FIo)/FIo.

Properties of the QDs-P4Ab conjugates
The luminescence and monodispersity of the QDs-P4Ab conjugates are essential for a QDbased fluorescence immunoassay to achieve high sensitivity and performance. In Fig. 1(a), gel images of the QDs-P4Ab conjugates are shown clearly under UV light. The fluorescence emission intensity of the QDs-P4Ab conjugates changed with different amounts of the antiprogesterone antibody (P4Ab) and QDs with fixed amounts of QDs (3.2 mg/ml) or P4Ab (250 ng/ml). In the gel image, the QDs-P4Ab conjugates moved slower than QDs alone because of the decrease in negative charge on the surface of QDs after conjugation with P4Ab. An adequate combination of the QDs and P4Ab for conjugation could be selected from the gel image with a high emission intensity at the band site, as shown in Fig. 1(a), where few QDs and P4Ab have been washed without conjugation. In this work, a combination of 1.6 mg/ml QDs and 250 ng/ml P4Ab could be used for the optimal conjugation of the QDs to P4Ab. After conjugation of the QDs to P4Ab via EDC/NHS coupling chemistry, the QDs-P4Ab conjugates were separated and purified from a number of components such as EDC, NHS, unconjugated QDs, and P4Ab by using ultrafilter-centrifugation with YM-100 and a fast protein liquid chromatography (FPLC) system equipped with a Sephadex G-100 gel column. Then, the QDs-P4Ab conjugates were concentrated with a freeze dryer and used either for immobilization onto a well in a direct binding assay or for affinity binding with the sample as a second antibody in a sandwich assay. As shown in Fig. 1(b), there were no differences in the fluorescence spectra between the CdSe/ZnS QDs and the purified QDs-P4Ab conjugates after conjugation. The excitation wavelength is in the range of 300-530 nm and the emission band edge is 590 nm.

Surface morphology of the APTMS sol-gel membranes
The surface morphologies of the APTMS sol-gel membranes in the absence and presence of the QDs-P4Ab conjugates were characterized by AFM and SEM, respectively, and they are shown in the images in Fig. 2. In the absence of QDs-P4Ab conjugates, the mean height (Ra, 0.333 nm) and the mean roughness (Rq, 0.404 nm) indicated that the APTMS sol-gel membrane had a flattened and smooth surface (Fig. 2(a)). The APTMS sol-gel membrane in the presence of QDs-P4Ab conjugates had a larger mean height (Ra, 1.410 nm) and a larger mean roughness (Rq, 1.701 nm) than the sol-gel membrane without the QDs-P4Ab conjugates ( Fig. 2(b)). SEM images at different scales were also taken to show the surface morphology of the APTMS sol-gel membrane in the presence of the QDs-P4Ab conjugates (Fig. 2(c) & 2(d)).

Development of QDs-based fluorescence immunoassays
Two heterogeneous immunoassays have been developed by immobilizing either the antiprogesterone antibody (P4Ab) or the QD-P4Ab conjugates onto a well surface of a 96-well microtiter plate via the APTMS sol-gel membrane. These two assays have been characterized for the determination of progesterone in human serum, and their results are compared with those of a commercially available progesterone assay method, RIA.

Sandwich immunoassay
The amount of P4Ab immobilized on the well via the APTMS sol-gel membrane affects the performance and sensitivity of a sandwich assay. Therefore, different amounts of P4Ab were immobilized on the well surface, and their effects on progesterone measurements are shown in Fig. 3. The normalized fluorescence intensity with progesterone concentrations was higher at 12.5 ng P4Ab/100 μl/well than 6.25 ng P4Ab/100 μl/well. Therefore, in this sandwich immunoassay, the concentration of 12.5 ng P4Ab/well led to immobilization on the well surface through the APTMS sol-gel membrane.   The performance of a sandwich immunoassay can be affected by the binding capability between the immobilized P4Ab and P4 as well as that between P4 and the QD-P4Ab conjugates. The binding capability is related to the incubation time, buffer pH value, amounts of immobilized P4Ab, etc. The effects of the binding between P4 and the QDs-P4Ab conjugates on the assay performance were investigated with different incubation times (3, 6, and 24 hours) at various P4 concentrations. Fig. 4 shows the normalized fluorescence intensity with different progesterone concentrations and at different incubation times. The incubation time between P4 and the QDs-P4Ab conjugates did not significantly affect the change in the normalized fluorescence intensity, i.e. the assay performance. Samples containing ascorbic acid (up to 1.0 mM), Na + ion (to 150 mM) or estrogen and testosterone were introduced into a well immobilized with P4Ab, then the QDs-P4Ab conjugates were added to the well. After incubating and washing the well with 10 mM PBS (pH 7.4), the normalized fluorescence intensity decreased for samples containing 0.1 mM ascorbic acid. This indicates that the presence of ascorbic acid in the samples interfered with the binding between P4 and P4Ab in the sandwich immunoassay. However, in the samples with other components, the normalized fluorescence intensity was not significantly changed (less than 10%) (data not shown).
The dissociation of the binding among the samples (P4), the immobilized P4Ab, and the QDs-P4Ab conjugates in the sandwich assay were investigated using three elution buffers, as in the direct binding assay. The changes in the normalized fluorescence intensity with different progesterone concentrations are shown in Fig. 5 with three elution buffers and elution times. When no elution buffer was used, the normalized fluorescence intensity increased with increasing progesterone concentration. The normalized fluorescence intensity stayed almost constant at 0.12, with 0.1 M glycine•HCl at an elution time of 5 or 10 minutes. The results show that 0.2 M NaOH did not dissociate the binding between P4 and the QDs-P4Ab conjugates at an elution time of 5 minutes or more. The 50 mM Tris-HCl buffer dissociated the binding very effectively at an elution time of 10 minutes, but at an elution time of 5 minutes, the normalized fluorescence intensity with 5,000 pg/ml of progesterone (P4) remained at about 0.23. This indicates that 0.1 M glycine•HCl can be used to dissociate the binding between P4 and the immobilized P4Ab in the sandwich immunoassay.  A number of other experimental conditions, including immobilization time, were also investigated and optimized. Table 1 summarizes the final experimental conditions that were ultimately used in the sandwich immunoassay. Under the final experimental conditions of the sandwich assay, the performance of the assay was assessed with a series of standard solutions with different progesterone concentrations. The calibration curve is shown in Fig. 6. The regression equation and parameters were obtained from a 4-parameter-logistic curve by using OriginPro Software (OriginLab Co., USA).
The plot indicates that linear concentration range of P4 was from 2184.6 to 117,082 pg/ml. The detection limit, which was calculated from 10 % of the maximum normalized fluorescence intensity (i.e., 3.28578), was is 553.9 pg/ml. The sensitivity of the assay was 18,251.96 pg/ml. The precision of the assay was estimated from the relative standard deviation (RSD) of the signals obtained for the analysis of a triplicate of samples containing 2,000 pg/ml, and it was ultimately found to be 3.7 %. The RSD in all points is <5%. Parameter A is 0.09517, which is the background of the normalized fluorescence intensity; parameter B is 3.28578, which is the maximum normalized fluorescence intensity obtained; parameter C is 18,251.96 pg/ml, which is the sensitivity of the assay; and parameter D is 0.72629, which is the slope of the curve. The regression coefficient (R 2 ) of the curve is 0.99235. Data are represented as mean+SD (n=3).

Direct binding immunoassay
The immobilization capacity of the QDs-P4AB conjugates onto the well surface affects the performance and sensitivity of the immunoassay. The immobilization of the QDs-P4AB conjugates can be accomplished in several ways: a) encapsulation of the QDs-P4Ab conjugates into the APTMS sol-gel membrane on the well surface, b) peptide binding between the carboxyl group of the water-soluble QDs in the QDs-P4Ab conjugates and the amine group of the APTMS in the APTMS sol-gel membrane etc.
The encapsulation of the QDs-P4Ab conjugates into the APTMS sol-gel membrane can be influenced by the gellation time of the APTMS sol-gel solution introduced into the well of a 96-well microtiter plate. The effects of the gellation time on the immobilization capacity of the QDs-P4Ab conjugates were investigated by spreading the sol-gel solution into wells and incubating the samples for 5, 10, 20, and 30 minutes at room temperature. The measurement results of the fluorescence intensity at various P4 concentrations showed that the intensity decreased as the gellation time of the sol-gel solution increased. That is, due to their soft texture, the QDs-P4Ab conjugates were more successfully encapsulated into the APTMS sol-gel membrane at a short gellation time (e.g. 5 min) than at a long gellation time (e.g., 30 min) (data not shown).
In this assay, the direct binding of P4 to the immobilized QDs-P4Ab conjugates decreased the fluorescence emission intensity of the QDs, because it changed the negative charge distribution around the QDs [28].
The long-term stability of the QDs-P4AB conjugates immobilized in a well was investigated at different P4 concentrations. After 7 weeks the fluorescence intensity of the well was mostly maintained, i.e., over 80% of its initial value (data not shown). This showed that the immobilization of the QDs-P4Ab conjugates into the APTMS sol-gel network created a special strong texture.
Different amounts of the QDs-P4Ab conjugates were immobilized onto a well surface and their effects on progesterone are shown in Fig. 7. There was no significant difference in the normalized fluorescence intensity in the wells with 50, 100, and 150 μl of the QDs-P4Ab conjugates. Therefore, in this direct binding assay, 100 μl of the QDs-P4Ab conjugates have been immobilized into the APTMS sol-gel membrane on a well surface. In an immunoassay, Biomolecular samples should be analyzed as quickly as possible. This is because the analysis time of a sample is affected by the binding between the antigen and antibody, and the binding of P4 to the immobilized QDs-P4Ab conjugates depends on the incubation time of P4. The change in the normalized fluorescence intensity with different incubation times and at various P4 concentrations is shown in Fig. 8. The normalized fluorescence intensity decreased in the concentration range from 0 to 5,000 pg/ml, regardless of the P4 incubation time. Next, the interfering effects of ascorbic acid (up to 1.0 mM) as well as Ca ++ , K + , and Na + ions (to 150 mM) on the binding between P4 and the QDs-P4Ab conjugates were investigated. The fluorescence intensity of the samples with additives did not change significantly compared to the intensity of the sample without additives. The addition of estrogen and testosterone to the sample did not interfere with the binding of P4 to the QDs-P4Ab conjugates. (data not shown). However, there may be some amino acids and proteins in samples which would interfere with the binding of the samples (P4) to the QDs-P4Ab conjugates.
The dissociation of the binding between sample (P4) and the QDs-P4Ab conjugates plays an important role in the repeated use of a well immobilized with the QDs-P4Ab conjugates. The effects of three elution buffers on the dissociation are shown in Fig. 9 at a few concentrations of progesterone. When no elution buffer was used, the normalized fluorescence intensity decreased with increasing concentrations of progesterone.
It was also decreased with 0.1 M glycine•HCl for an elution time of 5 or 10 minutes, where the binding between P4 and the QDs-P4Ab conjugates was not dissociated completely. However, the binding between P4 and the QDs-P4Ab conjugates was not eluted with 0.2 M NaOH. From these results, we used 0.1 M glycine•HCl to effectively dissociate the binding between P4 and the QDs-P4Ab conjugates immobilized in a well. To wash away the proteins in the samples and in some of the unbound samples, 10 mM PBS with surfactants (Tween X20 and X100) was employed after incubating the sample in a well immobilized with the QDs-P4Ab conjugates. The fluorescence intensity did not change significantly (less than 10%) with PBS containing surfactants between different progesterone concentrations.
Some other experimental conditions, such as sample volume, were investigated, and the final experimental conditions ultimately used for the assay are summarized in Table 2. Under the final experimental conditions, the performance of the direct binding assay was assessed as a series of standard solutions with different progesterone concentrations. The calibration curve was adjusted to an exponential decay curve [29], as shown in Fig. 10.
The regression equation and parameters were obtained from a 4-parameter-logistic curve by using OriginPro Software (OriginLab Co., USA). The plot indicates that the linear concentration range of P4 was from 28.95 to 26,607.7 pg/ml. The detection limit, which was calculated from 10% of the minimum normalized fluorescence intensity (i.e. -0.90743), was 3.320 pg/ml,. The sensitivity of the assay was 987.24 pg/ml. The precision of the assay was estimated from the RSD of the signals obtained for the analysis of a triplicate of samples containing 2,000 pg/ml, and it was ultimately found to be 4.5 %. Fig. 10. Calibration curve for P4 concentrations in the sandwich immunoassay. Each point on the curve is the average of three measurements. The error bars correspond to + 3 SD. The RSD in all points is <5%. Parameter A1 is -0.01391, which is the background of the normalized fluorescence intensity; parameter B1 is -0.90743, which is the maximum normalized fluorescence intensity obtained; parameter C1 is 987.24 pg/ml, which is the sensitivity of the assay; and parameter D1 is 0.41518, which is the slope of the curve. The regression coefficient (R 2 ) of the curve is 0.99709. Data are represented as mean+SD (n=3).

Applications to real samples
Two types of heterogeneous immunoassay, i.e., a direct binding assay and a sandwich assay, were applied to determine the P4 concentrations in several real samples of human serum, and these were compared with the results obtained using a commercial RIA kit. A calibration curve from a commercial RIA kit was constructed, and the expression was obtained using a least square procedure as follows: CPM = 407.24 + 2561.02 * exp (-0.015 * X) where CPM is the arbitrary unit for the gamma counter and X is the P4 concentration in the range from 0 to 10,000 pg/ml. The regression coefficient (R 2 ) of the curve was 0.9916. Fig. 11 shows the correlations of the results obtained from two heterogeneous immunoassays and the RIA method. Some samples (Nos. [7][8][9] were spiked by adding a standard P4 solution (5,000 pg/ml) to real samples. The mean difference in P4 concentrations determined using the three methods was within 0.15, i.e. 15.0 %, resulting from the measurement errors and regression equations of each calibration curve. That is, the standard deviation among the three assays was not high and was within the range of acceptability, i.e. less than 20%. This clearly indicates that the fluorescence heterogeneous immunoassays developed here may be comparable and acceptable alternative tools for the determination of P4 concentrations in clinical diagnosis.

Conclusion
Fluorescence heterogeneous immunoassays using CdSe/ZnS core/shell QDs were newly developed in this study to determine the concentrations of progesterone in human serum. The detection scheme was based on the direct change in the fluorescence intensity of QDs on the binding of the QDs-P4Ab conjugates to P4. Hydrophilic CdSe/ZnS QDs were conjugated to an anti-progesterone antibody (P4Ab) via EDC/NHS chemistry. Two types of heterogeneous assay were studied in a 96-well microtiter plate. While the direct binding assay was based on the direct binding of P4 to the QDs-P4Ab conjugates immobilized onto the APTMS sol-gel membrane of the wells, the sandwich assay was based on the binding of P4 to P4Ab immobilized onto the APTMS sol-gel membrane of the wells, and then on the binding of P4 to the QDs-P4Ab conjugates as a second antibody. The change in the fluorescence intensity of the QDs-P4Ab conjugates was correlated to the P4 concentrations, as the intensity increased with increasing P4 concentrations in the sandwich assay, but it decreased with increasing P4 concentrations in the direct binding assay. While the sandwich assay had a detection limit of 553.9 pg/ml and a sensitivity of 18,251.96 pg/ml with a linear range of 2184.6 -117,082 pg/ml, the direct binding assay had a detection limit of 3.32 pg/ml and a sensitivity of 987.24 pg/ml with a linear range of 28.95 -26,607.7 pg/ml. This indicates that the developed heterogeneous assays have been successfully applied for the determination of P4 in real human serum. The results showed a good correlation with the accredited RIA, suggesting that the developed assays meet the demands for clinical diagnosis.