Homogeneous Noncompetitive Immunoassay for 17β-Estradiol Based

Jun 30, 2007 - ... ochratoxin A based on intrinsic fluorescence of an antigen and antibody complex. Taihua Li , Ki-Seok Jeon , Yung Doug Suh , Min-Gon...
0 downloads 0 Views 200KB Size
Download PDF
Anal. Chem. 2007, 79, 5935-5940

Homogeneous Noncompetitive Immunoassay for 17β-Estradiol Based on Fluorescence Resonance Energy Transfer Tiina Kokko,* Leena Kokko, Timo Lo 1 vgren, and Tero Soukka

Department of Biotechnology, University of Turku, Tykisto¨katu 6A, 6th floor, FIN-20520 Turku, Finland

We previously presented a homogeneous noncompetitive assay principle based on quenching of the fluorescence of europium(III) chelate. This assay principle has now been applied to the measurement of 17β-estradiol (E2) using europium(III) chelate labeled estradiol specific antibody Fab fragment (Eu(III)-Fab) as a donor and E2 conjugated with nonfluorescent QSY21 dye as an acceptor. Fluorescence could be measured only from those Eu(III)-Fab that were bound to E2 of the sample, while the fluorescence of the Eu(III)-Fab that were not occupied by E2 were quenched by E2-QSY21 conjugates. The performance of the assay was tested both in buffer and in the presence of serum. Because approximately half of the Fabs were incapable of binding to E2, the maximum quenching efficiency was only 55%. The lowest limits of detection were 18 and 64 pM in buffer and serum-based calibrators, respectively. The highest measurable concentrations were 0.4 nM in buffer and 1 nM in serum. The presented noncompetitive assay principle requires only one binder, and it can be applied to other haptens as well, providing that a suitable antibody is available. 17β-Estradiol (E2) is the most important female sex hormone. It regulates the female menstrual cycle as well as maintenance of the bone mass and cardiovascular system for both genders. Normal E2 levels vary according to age and menstrual cycle. Many applications for E2 assays are used in diagnostics, for example, when monitoring hormonal infertility treatment, assessing ovarian function, or diagnosing tumors. It has been shown that increased amounts of E2 and its metabolites stimulate the proliferation of cancer cells. Therefore, the pharmaceutical industry uses E2 assays in high-throughput screening while searching for a proper inhibitor for E2 synthesis in cancer cells.2 Due to human activity, E2, and other estrogens, are also found in many aquatic environments from where it can be carried into the aquatic animals. This external E2 can cause several serious problems to water animals, for example, abnormal sexual development3 or induced vitelloge* Corresponding author, Tel: +358 2 333 8067. Fax: +358 2 333 8050. E-mail: [email protected]. (1) Lippert, C.; Seeger, H.; Mueck, A. O. Life Sci. 2003, 72, 877-883. (2) Kokko, L.; Johansson, N.; Lo¨vgren, T.; Soukka, T. J. Biomol. Screening 2005, 10, 348-354. (3) Guillette, L. J., Jr.; Gross, T. S.; Matter, J. M.; Percival, H. F. Environ. Health Perspect. 1994, 102, 680-689. 10.1021/ac070417o CCC: $37.00 Published on Web 06/30/2007

© 2007 American Chemical Society

nin production in male trout,4 which can be used as a biomarker for estrogen exposure. Thus, analyzing E2 from environmental samples is of growing importance. Due to the fact that E2 is a small molecule (Mw ∼272), only one antibody can bind to it simultaneously. Therefore, most of the developed immunoassays for E2 are competitive. In competitive assays, the analyte competes with labeled analyte analogue for the binding sites of the recognizing agent. Thus, signal to be measured is created by those antibodies that are bound to the labeled analyte analogue. Therefore, in competitive assays, increasing the amount of analyte decreases the obtained signal. Whereas, in noncompetitive assays, the signal is obtained from those antibodies that are bound to the sample analyte. Thus, in noncompetitive assays, increasing the amount of analyte also increases the obtained signal. Sensitivity can be defined as the smallest amount of analyte that generates a distinguishable difference in the signal from the background (zero amount of analyte).5,6 In competitive assays, the signal at zero concentration is high, whereas in noncompetitive assays, the signal at zero amount of analyte is small. When the concentration is slightly elevated from zero, only a small change in the signal occurs. It is typically easier to detect a small change from low signals than in high signals. Hence, the noncompetitive assays are capable of detecting smaller changes in analyte concentrations and are, therefore, considered to have better sensitivity than competitive assays.7 There are, however, some competitive assays with relatively good sensitivity that amplify the signal prior to the measurement. Zhao et al. developed a heterogeneous competitive chemiluminescence-based assay for wastewater E2, where the signal is amplified enzymatically. The detection limit of the assay was 5.4 pg/mL.8 Lamminma¨ki et al. designed a competitive immunoassay for E2, which utilized the enhancement of fluorescence of Eu(III) chelate. The obtained sensitivity was 8 pM.9 In homogeneous competitive assays, the intensity of the signal can be increased by using particle labels. One particle contains a large number of lanthanide chelates and thus enhances the intensity of fluores(4) Tanaka, T.; Takeda, H.; Ueki, F.; Obata, K.; Tajima, H.; Takeyama, H.; Goda, Y.; Fujimoto, S.; Matsunaga, S. J. Biotechnol. 2004, 108, 153-159. (5) Ekins, R. P. Clin. Chem. 1998, 44, 2015-2030. (6) Ekins, R. P.; Chu, F. W. Clin. Chem. 1991, 47, 1955-1967. (7) Saviranta, P. U.S. Patent 6,037,185, 2000. (8) Zhao, L.; Lin, J-M.; Li, Z.; Ying, X. Anal. Chim. Acta 2006, 778, 290-295. (9) Lamminma¨ki, U.; Westerlund-Karlsson, A.; Toivola, M.; Saviranta, P. Protein Sci. 2003, 12, 2549-2558.

Analytical Chemistry, Vol. 79, No. 15, August 1, 2007 5935

cence. For a recent review, see Steinkamp and Karst.10 Kokko et al. constructed an E2 assay using Fab-coated europium(III) nanoparticles as a donor and E2-Alexa680 conjugate as an acceptor. With this assay, 70 pM concentrations were detectable.11 Kuningas et al. designed an E2 assay where Fab-coated upconverting phosphorparticles acted as a donor and E2-oyster-556 was the acceptor. The obtained lowest limit of detection in assay buffer was 400 pM.12 We previously presented a principle for a homogeneous noncompetitive assay using streptavidin and biotin as a model.13 The purpose of this article is to present a homogeneous noncompetitive assay for E2 based on fluorescence resonance energy transfer (FRET) and compare its characteristics to FRET-based competitive assay. The assay requires only one antibody fragment (FabS16) labeled with intrinsically fluorescent europium(III) chelate and E2 analogue conjugated with nonfluorescent quencher dye. The performance of the assay was tested both with buffer and with serum-based calibrators. EXPERIMENTAL SECTION Measurement Buffer and Instruments. Buffer for all assays and dilutions containing 0.05 M Tris-HCl, pH 7.75, 0.9% (w/v) NaCl, 0.05% (w/v) NaN3, 0.01% (v/v) Tween 40, 0.05% (w/v) bovine γ-globulin, 20 µM diethylenetriaminepentaacetate, and 0.5% (w/ v) bovine serum albumin (BSA) was purchased from Innotrac Diagnostics (Turku, Finland). Fluorescence was measured using time-resolved detection with a Victor 1420 Multilabel counter from Wallac, Perkin-Elmer Life and Analytical Sciences (Wellesley, MA). Excitation wavelength was 340 nm, and measuring wavelength was 615 nm. Delay time and measurement time were both 400 µs. Sensitized emission of AlexaFluor680, generated by energy transfer, was measured at 730 nm with a modified 1234 Delfia Research Fluorometer (Wallac, Perkin-Elmer Life and Analytical Sciences) equipped with a 730nm band-pass emission filter with 10-nm bandwidth (Coherent Inc, Santa Clara, CA), a red-sensitive R2949 photomultiplier tube (Hamamatsu Photonics, Shimokanzo, Japan), and a 340-nm DUG11 excitation filter (Perkin-Elmer Life and Analytical Sciences) using 75-µs delay and 400-µs measurement time. All the washings and aspirations of the wells were done using 1296-026 Delfia Platewash. For the shakings of the plates, 1296003 Delfia Plateshake was used. Both of these instruments were acquired from Wallac, Perkin-Elmer Life and Analytical Sciences. BSA-Treated Microtitration Plates. All assays were done in Low Fluor 96-well Maxisorp microtitration plates purchased from Nunc (Roskilde, Denmark). Wells were treated with BSA prior to the assays to prevent nonspecific binding. Solution containing 0.1% (w/v) BSA (Bioreba), 0.1% (w/v) Germall II (ISP, Wayne, NJ), and 3% (w/v) trehalose (Sigma-Aldrich, St. Louis, MO) in 0.05 M Tris-HCl buffer, pH 7.2, was incubated for 1 h, at room temperature with low shaking. After incubation, plates were aspirated, dried, and stored in +4 °C until used. Europium(III) Chelate Labeled FabS16. E2-specific antibody FabS16 fragment was produced and purified as described (10) Steinkamp, T.; Karst, U. Anal. Bioanal. Chem. 2004, 380, 24-30. (11) Kokko, L.; Sandberg, K.; Lo ¨vgren, T.; Soukka, T. Anal. Chim. Acta 2004, 503, 155-162. (12) Kuningas, K.; Ukonaho, T.; Pa¨kkila¨, H.; Rantanen, T.; Rosenberg, J.; Lo ¨vgren, T.; Soukka, T. Anal. Chem. 2006, 78, 4690-4696. (13) Kokko, T.; Kokko. L.; Soukka, T.; Lo¨vgren, T. Anal. Chim. Acta. In press.

5936

Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

by Kokko.11 FabS16 was labeled with a long lifetime and intrinsically fluorescent 9-dentate europium(III) chelate (2,2′,2′′,2′′′-{[2(4-isothiocyanatophenyl)ethylimino]bis(methylene)bis{4-{[4-(Rgalactopyranoxy)phenyl]ethynyl}pyridine-6,2-diyl}bis(methylenenitrilo)}tetrakis(acetato)europium(III).14 For one labeling reaction, 160 µg of FabS16 and a 10-fold molar excess of europium chelate, dissolved in water, was used. pH was adjusted by adding 13 µL of 1 M carbonate buffer. The total volume of the reaction was adjusted to 260 µL with water. The reaction was incubated over night in room temperature. The next day, the reaction was purified with gel filtration using Sephadex G50 matrix (purchased from Pharmacia Biotech, GE Healthcare, Fairfield, CT). Buffer in the purification was TSA, pH 7.75, (0.05 M TrisHCl, 0.9% (w/v) NaCl, and 0.5% (w/v) NaN3, all purchased from Sigma-Aldrich). After the gel filtration, the presence of free europium(III) chelate in the purified fractions was tested. Fractions were diluted and added in to Nanosep 30K Omega filter Eppendorf tubes (Pall Life Science, New York, NY) and centrifuged for 10 min with 4600g. The amount of Eu(III) chelate in the filtrate was measured at 615 nm with a Victor 1420 Multilabel Counter. Dialysis was used to remove the remaining free europium. Dialysis was done in TSA buffer, pH 7.75, using a Cellu Sep T2 membrane with pore size of 6000-8000 Da from Membrane Filtration Products Inc. (Seguin, TX). The dialysis lasted for 4 days, and the buffer was changed every 24 h. After purification, the labeling degree of the product was specified. The amount of europium(III) chelate was measured by comparing the fluorescence of the labeling reaction to the fluorescence of the europium standard. Purified product was diluted in Delfia Enhancement Solution (purchased from Wallac, Perkin-Elmer Life and Analytical Sciences) and incubated for 30 min before measurement. The amount of FabS16 was calculated from the absorbance at 280 nm after subtracting the absorbance of europium(III) chelate at 280 nm. The labeling degree was calculated as europium concentration per FabS16 concentration. E2 Conjugations with Acceptor Dyes. Amino-modified E2 (6-oxoestradiol 6[O-(6-aminohexyl)oximel])15 was conjugated to QSY21 carboxylic acid, succinimidyl ester quencher dye and AlexaFluor680 carboxylic acid succinimidyl ester (both from Molecular Probes, Eugene, OR) to create an E2-acceptor conjugate. For each reaction, 2 mg of E2 amino derivative was dissolved in 75 µL of ethanol. Approximately 0.5 mg of acceptor dye was dissolved in 40 µL of N,N-dimethylformamide (SigmaAldrich). Estradiol derivative and acceptor were mixed, and 15 µL of 1 M carbonate buffer, pH 9.0, was added. Thereafter, the volume of the reaction was adjusted to 300 µL with water. Reactions were protected from light and incubated in +37 °C with slow rotation for 4 h. After incubation, the reactions were frozen until purification. Reactions were purified using an RP-HPLC technique (instrumentation form Thermo Electron Corp., Waltham, MA) using a Genesis C18 column from Jones Chromatography (Grace Vydac, Hesperia, CA). Gradient purification was done using 0.05 M triethylammonium acetate (TEAA) buffer (Fluka Biochemica, Steinheim, Switzerland) in water, and 0.05 M TEAA in acetonitrile (14) von Lode, P.; Rosenberg, J.; Pettersson, K.; Takalo, H. Anal. Chem. 2003, 75, 3193-3201. (15) Mikola, H.; Ha¨nninen, E. Bioconjugate Chem. 1992, 3, 182-186.

(from J.T. Baker, Phillipsburg, NJ) as buffers. The amount of acetonitrile was increased from 65 to 100% in 38 min. After purification, the amounts of E2 and acceptor dye were measured with absorbance. The appropriate wavelengths and molar absorbtivities for the acceptors were provided by the manufacturer. Homogeneous Competitive Assay. The competitive assay utilized Eu(III)-labeled FabS16 as a donor label and E2 conjugated with fluorescent AlexaFluor680 as an acceptor label. First 15 µL of assay buffer was added to BSA-treated wells. Thereafter, 5 µL of E2 standards from 0 to 500 nM with three replicas were added. Thus, the E2 concentrations in total well volume of 50 µL were 0-50 nM. Then 10 ng of Eu(III)-Fab in 15 µL was added. Wells were incubated for 30 min at room temperature and with low shake. After incubation, 15 µL of E2-AlexaFluor680 of 15 nM was added and incubation was continued for 15 min. After incubation, the sensitized emission of the AlexaFluor680 at 730 nm was measured using a modified 1234 Delfia Research Fluorometer. Homogeneous Noncompetitive Assay. The amount of E2QSY21 and the time required for sufficient binding of E2quencher to Fab was optimized. First, 50 µL of assay buffer with 0 or 2 µM E2 was added into BSA wells with three replicas. Thereafter, 10 ng of Eu(III)-FabS16 in 25 µL was added and wells were incubated for 30 min in room temperature with slow shaking. After incubation, 25 µL of E2-QSY21 between 0 and 1 µM was added and incubation was continued. Fluorescence of the Eu(III) chelate at 615 nm was measured after 5, 10, 15, 30, and 60 min of incubation with Victor 1420 Multilabel Counter. The homogeneous noncompetitive assay was done using Eu(III)-labeled FabS16 as a donor label and E2 conjugated with nonfluorescent QSY21 quencher dye as an acceptor label. First, 50-µL samples of E2 standards 0-4 nM (final concentrations in the total reaction volume of 100 µL were 0-2 nM), with three replicas, were added to BSA-treated wells. Then 8 ng of Eu(III)FabS16 in 25 µL was added. After 30 min of incubating the wells at room temperature with low shaking, 25 µL of 500 nM E2QSY21 was added and incubation was continued for 15 min. Thereafter, fluorescence of the europium(III) chelate at 615 nm was measured using a Victor 1420 Multilabel Counter. Homogeneous Noncompetitive Assay in the Presence of Human Serum. The functionality of the assay with serum samples was tested using male serum spiked with E2. Blood samples were let to settle at room temperature for 30 min. Thereafter, tubes were centrifuged (Eppendorf centrifuge 5810 R, Eppendorf, Hamburg, Germany) for 10 min with 1800g force. Separated serum was transferred to plastic tubes and frozen in -20 °C until used. For the standard curve, 40-µL serum samples were spiked with 0-6.25 µM E2 (final concentrations in the total well volume of 100 µL were 0-2.5 µM). Thereafter, 1.6 µL of 31.5 µM Mesterolone (1R-methylandrostan-17β-ol-3-one, Sigma-Aldrich) with a final concentration of 1 µM, was added. Serum calibrators were then incubated at room temperature for 1 h. The principle of the homogeneous assay in serum was similar to the homogeneous assay principle in assay buffer. First, 25 µL containing 4 ng of Eu(III)-FabS16 in assay buffer was added into BSA-treated wells. Thereafter, 10-µL samples of serum spiked with E2 (six replicas for serum with no E2 (background) and three

Figure 1. Standard curve, with standard error bars, for homogeneous competitive E2 assay. Lowest limit of detection is presented with dotted lines. CVs (b) are calculated from concentrations.

replicas for serum samples containing E2) were added. Addition of 40 µL of assay buffer increased the volume to 75 µL. Wells were then incubated at room temperature and with low shaking for 30 min. After incubation, 25 µL of 1 µM E2-QSY21 conjugate was added and incubation was continued for 15 min. Finally, fluorescence of the europium(III) chelate was measured using a Victor 1420 Multilabel Counter. RESULTS AND DISCUSSION Reagents. FabS16 was labeled with intrinsically fluorescent 9-dentate europium(III) chelate. The gel filtration was unable to completely separate the free europium from Eu(III)-Fab complex since some free europium(III) chelate was detected in the purified fractions. Therefore, fractions were dialyzed in order to remove the remaining free label. After dialysis, no free Eu(III) chelate was detected. The obtained labeling degree after dialysis was 1.7. The amino-modified E2 analogue was conjugated with fluorescent AlexaFluor680 acceptor and nonfluorescent QSY21 quencher. The HPLC-purified fractions of E2-acceptor dye conjugates were characterized with absorbance measurements. Fractions that contained similar amounts of both acceptor dye and E2 were selected. Homogeneous Competitive Assay. In order to test whether the noncompetitive assay has a lower limit of detection than the corresponding competitive assay, a homogeneous competitive assay utilizing the same binder as the noncompetitive assay was constructed. In the assay, E2 competes for the binding sites of E2-specific antibody Fab fragment labeled with europium(III) chelate with E2-AlexaFluor680 conjugate. Sensitized emission of AlexaFluor680 can be measured when the europium(III) chelate and AlexaFluor680 are in close enough proximity for FRET to occur. Thus, signal can be measured only from Eu(III)-Fab and E2-AlexaFluor680 complexes. The obtained standard curve for the assay is presented in Figure 1. The lowest limit of detection, calculated from the mean of the background subtracted with 3 times the standard deviation, was 1.2 nM (in total well volume of 50 µL). The highest measurable concentration was 10 nM, which equals 10% of the maximum signal. Thus, the practical range for the assay was 1.2-10 nM. Principle of the Homogeneous Noncompetitive Assay. The homogeneous noncompetitive assay for E2 was based on FRET Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

5937

Figure 2. Normalized emission spectrum of the europium(III) chelate (solid line) and the absorption spectra of the QSY21 (dotted line) and AlexaFluor680 (dashed line) acceptors.

Figure 4. Results for the optimizations of the homogeneous assay. Different incubation times: 5 (9), 10 (b), 15 (2), 30 (1), and 60 ([) min.

Optimization of the Homogeneous Noncompetitive Assay. The optimum amount of E2-QSY21 was tested with different amounts of E2-QSY21. It has been shown that 30 min is sufficient time for E2 to reach binding equilibrium to FabS16.11 Therefore, 30 min was used for the first incubation. The optimum length of the E2-QSY21 incubation was tested using 5-60-min incubation times. Results for these optimizations are presented in Figure 4. Quenching efficiencies were calculated using eq 1:

quenching efficiency ) max signal - signal of zero dose × 100% (1) max signal Figure 3. Different situations during measurement. (A) FabS16 labeled with intrinsically fluorescent europium(III) chelate emits at 615 nm when excitated at 340 nm. (B) When Eu-FabS16 is bound to estradiol of the sample emission is also observed at 615 nm. (C) When Eu-FabS16 is bound to estradiol-quencher conjugate, fluorescence of the europium(III) chelate is quenched.

between europium(III) chelate (donor) and nonfluorescent quencher dye (acceptor). QSY21 was selected for the quencher in this assay, since the absorption spectrum of QSY21 quencher dye overlaps well with the main emission peak (615 nm) of europium(III) chelate. The emission spectrum of the Eu(III) chelate and absorption spectra of QSY21 and AlexaFluor680 are presented in Figure 2. When the assay was performed, Eu(III)-FabS16 was, at first, incubated together with sample containing E2 whereupon Eu(III)-FabS16 bound to E2. Thereafter, E2-QSY21 conjugate was added. Since the dissociation of E2 from FabS16 is slow,9 there is very little competition between E2 and E2-QSY21 for the binding sites of the Fab. Thus, E2-quencher can bind only to those Eu(III)-FabS16 fragments that are not occupied by E2 of the sample. The different situations during measurement are presented in Figure 3. When europium(III) chelates are excited at 340 nm, FRET occurs between europium(III) chelate and quencher dye; thus, the fluorescence of Eu(III) chelate is quenched. Therefore, only the fluorescence of those Eu(III)-FabS16 fragments that are occupied by E2 of the sample can be measured at 615 nm. 5938

Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

The maximum signal of the assay was obtained when there was a high excess of free E2. Thus, all the binding sites of Eu(III)-FabS16 were occupied by E2 and no E2-QSY21 conjugate was able to bind to the Eu(III)-FabS16. Therefore, the emission of Eu(III) chelate was not quenched (situation B in Figure 3). The background signal ()signal of zero dose) was measured when there was no E2 in the well. Thus, all of the binding sites of Eu(III)-FabS16 were occupied by E2-QSY21 and maximum quenching was obtained (situation C in Figure 3). Based on these optimization results, 25 µL of 500 nM (final concentration being 125 nM in total well volume of 100 µL) E2QSY21 was chosen for the assay. When 500 nM E2-QSY21 was used, 15 min of incubation was sufficient for maximum quenching efficiency. The highest obtained quenching efficiency was only 55%. Thus, ∼45% of the emission of the Eu(III) chelate was not quenched. This was due to the fact that for some reason only approximately half of the Eu(III)-FabS16 was capable of binding E2 (results not shown). This applies to unlabeled FabS16 as well. It is unknown whether this problem relates to only this particular Fab or is it a universal problem of antibody Fab fragments. If this inactivity problem relates to only Fab fragments, it could be solved by using intact monoclonal antibodies. However, since the efficiency of the FRET depends strongly on the distance between the donor and the accepter, the distance between the quencher and the europium(III) chelate would most likely be too long for complete quenching of the fluorescence if intact antibody was used.

Figure 5. Standard curve, with standard error bars, for homogeneous noncompetitive E2 assay in assay buffer. Lowest limit of detection is illustrated with dotted lines. CVs (b) are calculated from concentrations.

Homogeneous Noncompetitive Assay In buffer. A standard curve for the homogeneous noncompetitive E2 assay was done using different amounts of E2 while the amount of Eu(III)-FabS16 was 8 ng/well and E2-QSY21 concentration was 125 nM. The obtained standard curve is presented in Figure 5. The equation of the linear part of the curve was: y ) 172000x1.02 (R2 ) 0.994). The lowest limit of detection, calculated as the concentration that equals the signal of 3× the standard deviation of the background, was 18 pM (4.9 pg/mL). Therefore, the practical range of the assay was 18 pM-0.4 nM. When this method was pretested using biotin and streptavidin as a model complex, the quadrivalent nature of the streptavidin deteriorated the linearity of the assay principle.13 As we speculated then, with monovalent binder (Eu(III)-FabS16) this problem did not exist. The exponent of the equation of the standard curve was 1.02; thus, the response to increasing E2 amounts was linear. The corresponding homogeneous competitive assay had a lowest limit of detection of 1.2 nM; thus, the noncompetitive assay has 67 times better detection limit and 2.5 times larger practical range than the competitive assay. The Fo¨rster radii when using Eu(III) chelate as a donor and AlexaFluor680 or QSY21 as an acceptor were calculated from the normalized emission and absorption data (Figure 2) as described by Selvin.16 The obtained Fo¨rster radii for AlexaFluor680 and QSY21 were 5.5 and 5.3 nm, respectively. Thus, the efficiency of energy transfer can be assumed to be similar in both assay principles. Both assays also used the same binder. Therefore, the improvement of the lowest limit of detection is most likely a consequence of the noncompetitive nature of the presented assay. E2-specific FabS16 has previously been used in other competitive homogeneous assay principles utilizing fluorescence resonance energy transfer. Both applications used particle labels as donors.11,12 The assay demonstrated in this article has better performance than the previously presented assays. The presented noncompetitive homogeneous assay also has performance equal to the heterogeneous competitive assay presented in the introduction,8 even though heteroge(16) Selvin, P. R. In Applied Fluorescence in Chemistry, Biology and Medicine; Rettig, W., Strehmel, B., Schrader, S., Seifert, H., Eds.; Springer-Verlag: Berlin, 1999.

Figure 6. Standard curve, with standard error bars, in the presence of human serum. Lowest limit of detection is indicated with dotted lines. CV profile (b) is calculated from concentrations.

neous assays are generally regarded as more sensitive than homogeneous assays.17 However, in order to obtain a low detection limit, the presented assay principle requires a high-quality antibody: The antibody has to have high affinity for the antigen and low dissociation rate to keep the antigen-antibody complex from separating during the incubation of the quencher-antigen. Assay Performance with Serum-Based Calibrators. The usability of the homogeneous noncompetitive assay principle with serum samples was tested using E2 spiked male serum. The presence of endogenous streroid binding proteins in serum impairs the usability of E2 assays with serum samples.18-20 To improve the assay performance and to block the endogenous steroid binding proteins, Mesterolone was added in to the E2spiked serum samples and only 4 ng of Eu(III)-FabS16 was used per well. The Mesterolone decreased dispersion and improved reproducibility, but it did not improve the lowest limit of detection of the assay (results not shown). The obtained standard curve for 10% serum of the total well volume is presented in Figure 6. The equation for the linear section of the curve was y ) 38000x1.03 (R2 ) 0.963). Even though the Mesterolone was added to the samples, the CV profile and the standard error bars are still rather large. The lowest limit of detection for the assay was calculated as the concentration that equals the signal of 3× standard deviation of the background and was 64 pM. The practical range of the assay, with presented conditions, was 64 pM-1 nM. As expected, the steroid binding proteins interfered with the assay by hindering the E2 binding to Fab fragments. Thus, the assay performance was poorer with serum than with buffer-based E2 calibrators. This is reported also in other homogeneous E2 assays.12 The lowest limit of detection increased ∼3.5 times when 10% serum of the total well volume was added. The effect of serum to FRET was tested by adding 0-50% serum into the wells shortly before the measurement. Thus, the serum could only affect the FRET, not the binding of E2 to Fab. The obtained quenching (17) Blomberg, K.; Hurskainen, P.; Hemmila¨, I. Clin. Chem. 1999, 45, 855861. (18) Brock, P.; Eldred, E. W.; Woiszwillo, J. E.; Doran, M.; Schoemaker, H. J. Clin. Chem. 1978, 24, 1595-1598. (19) Slaats, E. H.; Kennedy, J. C.; Kruijswijk, H. Clin. Chem. 1987, 33, 300302. (20) Masters, A. M.; Ha¨hnel, R. Clin. Chem. 1989, 35, 979-984.

Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

5939

efficiencies were equal with different serum amounts. Thus, the serum does not affect the FRET between europium(III) chelate and QSY21. Therefore, the presented assay principle can be used with serum samples as well.

required in order to be able to separate the Fab fragments capable of binding to E2 from the incapable ones. This separation would lower the background of the assay and enable the creation of a highly sensitive homogeneous noncompetitive assay.

CONCLUSIONS In this study, a noncompetitive homogeneous immunoassay for E2 based on FRET between europium(III) chelate (donor) and nonfluorescent quencher dye (acceptor) was demonstrated. The assay had significantly lower limit of detection than a FRET-based competitive homogeneous immunoassay using the same binder, even though the performance of the assay was somewhat impaired due to the inactive antibody Fab fragment. Thus, further work is

ACKNOWLEDGMENT This study was supported by TEKES, the Finnish Funding Agency for Technology and Innovation.

5940

Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

Received for review March 1, 2007. Accepted May 25, 2007. AC070417O