Homogeneous immunoassays based on fluorescence emission intensity variations of zinc selenide quantum dot sensors

Biosensors and Bioelectronics 41 (2013) 143–149

Contents lists available at SciVerse ScienceDirect

Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios

Homogeneous immunoassays based on fluorescence emission intensity variations of zinc selenide quantum dot sensors Jun Wang, T.J. Mountziaris n Department of Chemical Engineering, University of Massachusetts, Amherst, MA 01003, USA

a r t i c l e i n f o

abstract

Article history: Received 8 May 2012 Received in revised form 1 August 2012 Accepted 1 August 2012 Available online 14 August 2012

The fluorescence emission intensity of ZnSe quantum dots (QDs) conjugated to proteins to form QD-based biomolecular sensors increases significantly upon binding of the sensors to target proteins in solution. This phenomenon enables the development of homogeneous, separation-free immunoassays for rapid quantitative detection of proteins in solution. Proof-of-principle assays were developed by dosing a solution containing a biomolecular target with a solution containing the corresponding QD-based sensor and monitoring the changes in the peak fluorescence emission intensity of the QDs. Direct immunoassays for detecting basic fibroblast growth factor (bFGF) and prostate-specific antigen (PSA) in solution were demonstrated using QD-anti-bFGF and QD-anti-PSA sensors. A competitive immunoassay for detecting human serum albumin (HSA) was also demonstrated by dosing samples containing HSA with QD-HSA sensors and free anti-HSA antibodies. The QD-HSA sensors were tested in 1000  diluted human serum and found to be unaffected by interference from other proteins. The lower limit of detection of the assays was equal to the lowest sensor concentration in the solution that can be unambiguously detected, typically less than 1 nM. The dynamic range of the assays was determined by identifying the sensor concentration above which optical interference between QDs affected adversely the observed fluorescence emission intensity. The upper limit of this concentration was 2.5 mM for 4 nm QDs. The ZnSe QD-based sensors were stable and preserved  80% of their initial peak emission intensity after two months in refrigerated storage. These biosensors have potential applications in rapid sensing of target proteins for emergency and point-of-care diagnostic applications. & 2012 Elsevier B.V. All rights reserved.

Keywords: Zinc selenide quantum dots Homogeneous separation-free immunoassay Fluorescence spectroscopy Optical biosensor Fluorescence emission intensity variations Rapid quantitative protein detection Point of care diagnostics Emergency care diagnostics

1. Introduction Semiconductor nanocrystals or quantum dots (QDs) are nanometer-sized inorganic crystals that have unique optical properties due to confinement of electron-hole pairs (excitons) by the grain boundary of the nanocrystals (Alivisatos, 1996; Brus, 1991). These include size-tunable photoluminescence, high quantum yields and molar extinction coefficients, and high photostability (Murray et al., 2000). QDs have attracted significant attention as fluorescent labels of proteins and cells for in vitro and in vivo imaging and for biological sensing applications (Michalet et al., 2005; Smith et al., 2006). The excellent photochemical stability and high brightness of QDs can increase detection sensitivity in immunoassays and their narrow and tunable emission spectra enable multiplexing (Goldman et al., 2005a, 2006; Sapsford et al., 2006). A variety of targets have been detected using QDs as fluorescent labels, including small molecules, protein disease markers, bacteria, and viruses (Gill et al., 2008; Sapsford et al., 2006). For example, QDs

n

Corresponding author. Tel.: þ1 4135456145; fax: þ1 4135453540. E-mail address: [email protected] (T.J. Mountziaris).

0956-5663/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bios.2012.08.003

have been employed in ELISA-type immunoassays (Wang et al., 2002; Woodbury et al., 2002), in Western blot analysis of proteins (Bakalova et al., 2005; Chen et al., 2009), in sandwich immunoassays that can detect prostate-specific antigen (PSA) (Kerman et al., 2007), in multiplexed sandwich immunoassays that can simultaneously detect four toxins (Goldman et al., 2004), and as donors for detection ¨ of small molecules and biological targets based on Forster Resonance Energy Transfer (FRET) between the QD and another fluorophore (Algar and Krull, 2008; Clapp et al., 2005; Goldman et al., 2005b; Medintz et al., 2003). However, the substitution of fluorescent proteins by QDs in biological sensing applications is not always practical due to the high cost of QDs and the high toxicity of typical CdSe-based QDs. In this paper we demonstrate the development of novel homogeneous (separation-free) assays that enable direct detection of target analytes by monitoring the variations of the peak emission intensity of QD-labeled biomolecular sensors upon binding to specific target analytes. The variations in the peak emission intensity of the QDs are caused by surface-induced electronic perturbations (Cadars et al., 2009). The detection of a target analyte by these assays is accomplished without employing a second fluorophore and without immobilizing the probe, target, or probe-target complex

144

J. Wang, T.J. Mountziaris / Biosensors and Bioelectronics 41 (2013) 143–149

onto the surface of a micro-well or micro-bead. This minimizes mass transport limitations that can dramatically decrease the response time of a sensor (Sikavitsas et al., 2002). Homogeneous assays that are robust and reliable can increase the speed and simplicity of detection (Kricka, 1994). These assays are typically designed to provide immediate response in the form of an optical or electrical signal when the specific target is detected, which is an important requirement for assays suitable for point-of-care or emergency care applications (Ligler, 2009). A potential limitation of homogeneous, separation-free assays is their reduced sensitivity due to interference of other molecules in the sample with the measured signal. Homogeneous assays were developed using ZnSe QDs as responsive fluorescent beacons that are conjugated with a probe biomolecule. ZnSe QDs are less toxic when compared to Cd-based ones and allow more environmentally friendly processing and disposal. We demonstrate direct homogeneous assays for rapid quantitative detection of basic fibroblast growth factor (bFGF) and prostate specific antigen (PSA) targets in phosphate buffer saline (PBS) solution using ZnSe QD-anti-bFGF and ZnSe QD-anti-PSA sensors, respectively. A competitive homogeneous assay for detecting human serum albumin (HSA) in PBS and diluted human serum was also developed by dosing the HSA-containing solution with ZnSe QD-labeled HSA and anti-HSA antibody. In contrast with typical heterogeneous assays that require multiple steps and immobilization of the target and sensor on a substrate surface before measuring the signal, these assays were much simpler and faster to execute, making them attractive for applications requiring rapid quantitative detection of biological targets.

2. Materials and methods 2.1. Materials Trioctylphosphine (TOP, 90%), selenium powder (  100 mesh, 99.5%), 1-hexadecylamine (HDA, 98%), diethylzinc (Et2Zn, 1.0 M solution in heptane), 11-mercaptoundecanoic acid (MUA, 95%), potassium tert-butoxide (95%), N-(3-Dimethylaminopropyl)-N0 ethylcarbodiimide hydrochloride (EDC), and N-hydroxylsulfosuccinimide (Sulfo-NHS) were purchased from Sigma-Aldrich. Methanol, ethylether, butanol, and hexane were purchased from Fisher. 10  Phosphate buffered saline (PBS) was purchased from Gibco. Human serum albumin (HSA, 99%, MW 67.6 kDa), monoclonal anti-human serum albumin antibody produced in rabbit (anti-HSA, MW 150 kDa), human serum from platelet poor human plasma (sterile-filtered, mycoplasma tested and virus tested), polyclonal anti-prostate specific antigen antibody (anti-PSA, MW 150 kDa) produced in rabbit, and prostate specific antigen (PSA, MW 28.4 kDa) from human semen were purchased from Sigma-Aldrich. Recombinant human basic fibroblast growth factor (bFGF, MW 16 kDa) and mouse anti-human basic fibroblast growth factor (anti-bFGF, MW 150 kDa) were purchased from Biolegend.

2.2. Synthesis and functionalization of ZnSe QDs ZnSe QDs were synthesized by injecting a mixture of diethylzinc and Se powder dispersed in TOP into hot HDA at 310 1C (Hines and Guyot-Sionnest, 1998). Samples were drawn at specific time intervals and the ZnSe QDs contained in them were characterized to determine their size and optical properties. Samples containing ZnSe QDs with sizes between 3 and 5 nm were used for the biological sensing experiments. These QDs were modified with MUA according to published protocols (Chan and Nie, 1998) and were dispersed in aqueous solutions.

2.3. Preparation of ZnSe QD-antibody and QD-antigen sensors A standard stock solution of HSA was prepared by directly dissolving 10 mg HSA into 1 mL of 1  PBS buffer. Anti-HSA, bFGF and anti-bFGF stock solutions were prepared by dissolving 100 mg of each one of the as-purchased proteins in 1 mL of sterile 18 MOcm water. As-purchased solutions containing 1 mg anti-PSA/mL and 2.83 mg PSA/mL were diluted with PBS buffer to prepare stock solutions of the two proteins. All protein solutions were filtered through a 0.20 mm filter into sterile glass vials and stored at 4 1C. ZnSe QDs functionalized with MUA were conjugated with proteins by a two-step procedure using EDC and Sulfo-NHS as cross-linkers, following a standard protocol (Hermanson, 1996). For example, 8 mg of EDC and 40 mL of freshly-made NHS-water solution containing 10 mg NHS/mL were added to 2.0 mL of QD PBS buffer solution containing 1.06  1015 particles/mL which is equivalent to a QD concentration of 1.75 mM. The solution was stirred slowly for 30 min at room temperature before adding 3 to 4 nmol of HSA. The ratio of the number of ZnSe QDs to the number of biological molecules available for conjugation was controlled to be close to 1:1. The mixture was incubated at room temperature for 4 h to complete the covalent coupling reaction and allowed to stay overnight to enhance QD stabilization.

2.4. Immunoassay Calibration lines for the immunoassays were established by measuring the peak emission intensity of the free sensor and the equimolar sensor þtarget complex as a function of sensor concentration. Two sets of 15 vials were prepared, each containing a progressively increasing concentration of sensor. One set consisted of fifteen 1 mL samples of sensor solution, each containing a progressively increasing amount of sensor. This reference set was used to generate the sensor-only calibration line. The second set was used to generate the equimolar sensor þtarget calibration line as follows: 15 vials were prepared, each containing less than 1 mL of sensor solution with a progressively increasing concentration. A small volume of a concentrated solution containing free antibody (or antigen) was added to each vial to deliver an amount of antibody (or antigen) that was equal to the amount of sensor contained in the vial. PBS buffer was subsequently added to make the final volume of the mixture in each vial to be equal to 1 mL. After a short mixing period on an orbital shaker, the fluorescence spectrum of each sample was measured and the peak emission intensity was plotted as function of the concentration of sensor in the final 1 mL mixture. The sensor-only calibration line provided the lower limit of fluorescence intensity during assay execution and the equimolar sensorþtarget line the upper limit. Each immunoassay was executed manually by dosing a sample solution containing the target with a solution containing the corresponding sensor. For direct immunoassays, the sensor solution contained a QD-labeled antibody. For competitive immunoassays, two dosing solutions were used, one containing QD-labeled antigen and the other free antibody. For the assay examples discussed in this paper, a series of samples with volume less than 1 mL was prepared, each containing the same amount of target. An increasing amount of QD-labeled antibody (for direct immunoassays) or labeled antigen and free antibody solution (for competitive immunoassays) was added into each vial, always starting with an amount that is smaller than the amount of target to be detected. The final volume in each vial was adjusted to be equal to 1 mL by adding PBS buffer. After mixing, the fluorescence spectrum of each solution was recorded and the peak emission intensity was plotted as a function of the concentration of the QD-labeled sensor that was added to each sample. For commercial applications, such an immunoassay can be

J. Wang, T.J. Mountziaris / Biosensors and Bioelectronics 41 (2013) 143–149

executed by adding progressively larger amounts of a sensor solution directly into a single sample. 2.5. Fluorescence spectroscopy The fluorescence emission spectra of ZnSe QDs were measured at room temperature using a Flurolog-3 Spectrofluorometer (Horiba Jobin Yvon) with the excitation wavelength set at 340 nm. The peak fluorescence emission wavelength of the QDs was between 390 and 420 nm corresponding to an average particle size between 2.7 and 4.6 nm (Mei et al., 2008).

3. Results and discussion ZnSe QDs, functionalized with MUA and dispersed in PBS buffer solution, were conjugated with probe proteins using EDC as the cross-linker. In all cases the fluorescence emission intensity of the QDs increased after conjugation with the biomolecules. An example is shown in Fig. 1, where the fluorescence emission spectra of ZnSe QDs are plotted for three samples consisting of: (1) MUA-capped ZnSe QDs, (2) QDs from the same population conjugated with anti-HSA, and (3) QDs from the same population conjugated with HSA. The conjugated QDs were prepared using a ratio of QDs to protein close to 1:1. The fluorescence emission

14 12

(3)

(2)

10 (1)

8 6 4 2 0 360

410 460 Wavelength (nm)

Fig. 1. Fluorescence emission spectra of (1) MUA-capped ZnSe QDs, (2) QD-antiHSA conjugates, and (3) QD-HSA conjugates, in PBS solution. All QDs were from the same population.

10 8 6

(1)

4 2 0 360

14

4 (2)

410 460 Wavelength (nm)

Fluorescence Emission (× 106, a.u.)

Fluorescence Emission (× 106, a.u.)

12

intensity of the QDs increased after conjugation with HSA and antiHSA, without any apparent change in the peak emission wavelength. In this case, the conjugation of the QDs with the antigen, HSA, results in the strongest emission intensity. The MUA-capped QDs are negatively charged and preserve  50% of their initial fluorescence emission intensity after 7 days in storage at room temperature. The conjugation of the MUA-capped QDs with negatively charged proteins, such as the ones used in this study, induces electronic perturbations in the QDs resulting in increased fluorescence emission intensity. In Fig. 2, the fluorescence emission spectra of three sensors and their corresponding binding proteins are plotted. The sensors were: QD-HSA (Fig. 2a), QD-anti-bFGF (Fig. 2b) and QD-anti-PSA (Fig. 2c). Aqueous solutions of these sensors were mixed with stoichiometric amounts of target proteins, i.e., anti-HSA, bFGF, and PSA, respectively, and the fluorescence spectra of the QDs were recorded before and after binding of each sensor to the target protein. In all cases, the fluorescence emission intensity of the QDs increased after binding of the sensor to the target and there was no apparent shift in the peak emission wavelength. This increase in the peak emission intensity of the QD-based sensors was used as the detection signal in homogeneous assays that enable rapid quantitative detection of biomolecular targets in solution. Since the detection protocol of the assay is based on the difference between the peak emission intensity of the sensor and the sensorþtarget complex, the most suitable sensor and assay type (direct or competitive) is selected by comparing the difference in the peak emission intensity between the sensor and sensorþtarget complex. The fluorescence intensity amplification that is observed when the biosensor binds its target depends on the negative charge distribution on the surface of the QD before and after binding. The charge distribution depends on the conformation of the biomolecules attached to the surface of the QD. It is prudent to investigate both types of sensors, the QD-labeled antibody for direct assays and QD-labeled antigen for competitive assays, in order to identify the sensor that maximizes the fluorescence amplification factor, thus maximizing the sensitivity and range of detection of the assay. For example, QD-HSA sensors and a competitive assay protocol were found to provide the best performance for HSA detection, whereas QD-anti-bFGF and QD-anti-PSA sensors and direct assays were selected for detection of bFGF and PSA, respectively. The conjugation of QDs with probe biomolecules and the binding of the QD-based sensors to their intended targets alters the negative charge distribution on the surface of the QDs. This perturbation of the surface charge alters the electronic structure of the QDs and enhances radiative electron-hole recombination that results in the observed increase in fluorescence emission intensity.

Fluorescence Emission (× 106, a.u.)

Fluorescence Emission (× 106, a.u.)

16

(2)

3

2 (1)

1

0 370

145

420 Wavelength (nm)

470

(2)

12 10 8

(1)

6 4 2 0 370

420 Wavelength (nm)

470

Fig. 2. Fluorescence emission spectra of ZnSe QD-biomolecular sensors and the corresponding sensor þtarget complex: (a) (1) ZnSe QD-HSA, (2) ZnSe QD-HSA/anti-HSA; (b) (1) ZnSe QD-anti-bFGF, (2) ZnSe QD-anti-bFGF/bFGF and (c) (1) ZnSe QD-anti-PSA, (2) ZnSe QD-anti-PSA/PSA.

146

J. Wang, T.J. Mountziaris / Biosensors and Bioelectronics 41 (2013) 143–149

It has been shown that surface-induced electronic perturbations can propagate six to seven atomic layers into ZnSe nanocrystals of sizes between 5 and 9 nm (Cadars et al., 2009). Structuredependent amplification of the fluorescence emission intensity has also been observed upon conjugation of ZnSe QDs with singlestranded (ss) DNA at 1:1 QD to DNA ratio and hybridization of the QD-ssDNA conjugates with free complementary ssDNA molecules (Wang et al., 2012). Direct homogeneous immunoassays that can be used to detect free bFGF and PSA in solution were developed by using QD-antibFGF and QD-anti-PSA as sensors, respectively. A schematic representation of such a direct assay is shown in Fig. 3(a). The assay employs two calibration lines, one corresponding to the free QD-based sensor and the other to the sensorþ target complex. Such an assay can be executed by dosing the solution containing the target with progressively increasing amounts of sensor, starting with the amount corresponding to the smallest detectable concentration of sensor in the target solution. Figs. 3(b) and (c) show results from a proof-of-principle direct assay that detects free bFGF in PBS. The assay was executed manually by adding progressively increasing amounts of a QDanti-bFGF sensor to a series of samples containing 0.14 mg (6.62 pmol) bFGF in PBS. Fig. 3(b) shows the two calibration lines corresponding to the sensor only and sensor þtarget complex, as well as the evolution of the peak fluorescence emission intensity of the solution during assay execution. The recorded peak emission intensity initially followed the sensorþtarget calibration line, indicating that free target was present in the sample. When the free target was depleted from the sample, the recorded peak emission intensity departed from the sensor þtarget line and transitioned towards the free sensor line. The point of departure from the sensor þtarget line corresponds to the initial concentration of free bFGF in the solution. At the point of departure from the sensor þtarget line all free bFGF is bound to the sensor and the

concentration of ZnSe QD-anti-bFGF sensor in the 1 mL solution resulting from mixing the initial sample with the sensor solution dose is 6.66 nM. Assuming ideal binding conditions and 1:1 binding of target to sensor, the amount of target bFGF contained in the sample is estimated to be 6.66 pmol. This is reasonably close to the amount of bFGF contained in the initial sample, which was 6.62 pmol. The recorded peak emission intensity of the solution during assay execution can be normalized for convenience using the formula: In ¼(I–Is)/(Ist–Is), where In is the normalized peak emission intensity, I is the measured peak emission intensity, Is is the peak emission intensity of the free sensor at the concentration corresponding to the total amount that has been already added to the sample, and Ist is the peak emission intensity corresponding to the sensor þtarget complex at the same concentration. When this scaling is applied the normalized peak emission intensity of a bound sensor becomes equal to unity and that of a free sensor becomes equal to zero. Fig. 3(c) shows the normalized emission intensity of the homogeneous assay plotted in Fig. 3(b). The calibration lines in this case are a horizontal line at In ¼1, corresponding to the sensorþtarget complex, and a horizontal line at In ¼0, corresponding to the free sensor. If all sensor molecules are bound to the target the normalized emission intensity has a value of 1. This value starts transitioning towards 0 when all free target molecules are bound and unbound sensor molecules are available in the solution. The point of departure from the sensorþtarget calibration line corresponds to the initial concentration of free target in the solution. Fig. 4 shows results from a homogeneous assay that employs ZnSe QD-anti-PSA sensors to detect free PSA in PBS. Fig. 4(a) shows data from assay calibration and execution. The two straight lines in the plot correspond to sensor þtarget (upper) and sensor only (lower). The assay initially follows the sensor þtarget line and departs from it when the free target is depleted. Fig. 4(b) shows the corresponding normalized data. For the example shown in Fig. 4, a solution containing 78.6 ng (2.75 pmol) of free

Bioconjugation

Antigen-Antibody binding

Sample from Patient

Peak Emission Intensity

Labeled Antibody

Sensor+Target

Sensor only

Concentration of Target

Sensor Concentration (labeled antibody)

1.2 Normalized Peak Emission Intensity

Peak Emission Intensity (× 106, a.u.)

28 QD-anti-bFGF QD-anti-bFGF/bFGF Immunossay

24 20 16 12 8 4 0

0

0.005

0.01

0.015

0.02

0.025

Concentration of Labeled anti-bFGF (µM)

1 0.8 QD-anti-bFGF QD-anti-bFGF/bFGF Immunoassay

0.6 0.4 0.2 0

0

0.005

0.01

0.015

0.02

0.025

Concentration of Labeled anti-bFGF (µM)

Fig. 3. Homogeneous direct immunoassay that detects free bFGF in PBS: (a) immunoassay schematic; (b) peak fluorescence emission intensity obtained by dosing the sample with progressively increasing amounts of ZnSe QD-anti-bFGF sensor and (c) normalized peak emission intensity.

J. Wang, T.J. Mountziaris / Biosensors and Bioelectronics 41 (2013) 143–149

1.2 QD-anti-PSA QD-anti-PSA/PSA Immunoassay

25

Normalized Peak Emission Intensity

Peak Emission Intensity (× 106, a.u.)

30

20 15 10 5 0

147

0

0.005

0.01

1 0.8 0.6 0.4 QD-anti-PSA QD-anti-PSA/PSA Immunoassay

0.2 0

0.015

0

Concentration of Labeled anti-PSA (µM)

0.002

0.004

0.006

0.008

0.01

Concentration of Labeled anti-PSA (µM)

Fig. 4. Homogeneous direct immunoassay that detects free PSA in PBS: (a) peak fluorescence emission intensity obtained by dosing the sample with progressively increasing amounts of ZnSe QD-anti-PSA sensor and (b) normalized peak emission intensity.

Peak Emission Intensity

Sensor+ antibody

25 20 15 10 5 0

Used to Calculate the Unknown Target Concentration

1.2 QD-HSA QD-HSA/anti-HSA Immunoassay

Normalized Peak Emission Intensity

Peak Emission Intensity (× 106, a.u.)

30

Sensor only

0

0.02

0.04

0.06

0.08

Concentration of Labeled HSA (µM)

1 0.8 0.6 0.4 QD-HSA QD-HSA/anti-HSA Immunoassay

0.2 0

0

0.005 0.01 0.015 0.02 0.025 0.03 Concentration of Labeled HSA (µM)

Fig. 5. Homogeneous competitive immunoassay that detects free HSA in PBS: (a) immunoassay schematic; (b) peak fluorescence emission intensity obtained by dosing the sample with progressively increasing amounts of ZnSe QD-HSA sensor and anti-HSA antibody and (c) normalized peak emission intensity.

PSA target in PBS buffer was used. At the point of departure from the sensorþtarget line, all free PSA is bound to the sensor and the concentration of ZnSe QD-anti-PSA sensor in the 1 mL solution resulting from mixing the initial sample with the sensor solution dose is 2.6 nM. Assuming ideal binding conditions and 1:1 binding of target to sensor, the amount of target PSA contained in the sample solution is estimated to be 2.6 pmol, or 74 ng. This is 5.8% lower than the amount of 78.6 ng PSA contained in the initial sample. A competitive assay that detects HSA in PBS and in diluted human serum was also performed as a proof-of-principle example. A schematic representation of the assay is shown in Fig. 5(a). In this case, the sensor was ZnSe QD-HSA with 1:1 protein to QD

ratio. The following procedure was followed to manually execute this assay: 1. Two concentrated stock solutions, one containing ZnSe QD-HSA (labeled antigen) and the second containing anti-HSA antibody, were prepared. 2. Two calibration lines were developed that corresponded to peak fluorescence emission intensity vs. molar concentration of sensor, for ZnSe QD-HSA (sensor only) and ZnSe QD-HSA/ anti-HSA bound complex (sensor þantibody). 3. The solution containing the unknown amount of free HSA was dosed with progressively increasing amounts of ZnSe QD-HSA and anti-HSA, starting with the smallest detectable amount of

J. Wang, T.J. Mountziaris / Biosensors and Bioelectronics 41 (2013) 143–149

ZnSe QD-HSA. The molar amount of anti-HSA added with each dose was initially equal to and subsequently slightly higher than the molar amount of ZnSe QD-HSA being added. 4. The ZnSe QD fluorescence emission from the sample was recorded and the peak emission intensity was plotted as function of the ZnSe QD-HSA concentration added to the sample. In competitive assays, when the sample does not contain any free target the recorded peak fluorescence intensity simply follows the upper calibration line corresponding to the peak fluorescence emission intensity of the equimolar sensor þantibody complex. If there is free target present in the sample at a concentration higher than the initial dose of sensor, the free target will compete effectively for the added antibody and most of the sensor molecules will not be able to bind the antibody. Under such conditions, the peak emission intensity recorded from the sample will initially follow the sensor-only lower calibration line. As more doses of sensor and antibody are added to the mixture, the free target antigen is progressively bound to the antibody and the concentration of the sensorþantibody complex increases, thus leading to a transition of the recorded fluorescence peak emission intensity towards the sensorþantibody upper calibration line. When the peak emission intensity first reaches the sensorþ antibody line, all sensor molecules are bound to the antibody and the total amount of antibody that was added to the sample up to that point is equal to the total amount of bound antigen (sensor plus target). The initial molar amount of free antigen (target) in the sample is estimated by computing the difference between the total molar amounts of antibody and sensor that were added to the solution up to that point. To demonstrate this procedure, progressively increasing amounts of ZnSe QD-HSA and anti-HSA were added to a set of vials containing a fixed amount of free HSA in PBS. The fluorescence emission intensity of the samples was measured and the results were plotted in Figs. 5(b) and (c). Fig. 5(b) shows data from the assay that initially follows the lower calibration line (sensor only) and transitions to the upper calibration line (sensorþantibody) when the free HSA is depleted from the sample. The corresponding normalized emission intensity plot is shown in Fig. 5(c). For this proof-of-principle assay, a solution containing 0.17 mg (2.634 pmol) of free HSA was used. When the recorded peak emission intensity reached the upper line, the entire amount of sensor and free HSA target were assumed to be bound to the added antibody. The initial amount of free HSA present in the sample was calculated by subtracting the total molar amount of the sensor from the total molar amount of antibody added to the sample. After executing the assay, it was found that 10 mL of 2.631 mM ZnSe QD-HSA (1:1) sensor solution and 35 mL of 0.826 mM anti-HSA solution had been added to the target solution when the peak emission intensity reached the upper calibration line. The difference between the anti-HSA and labeled HSA added to the target solution was 2.6 pmol, a value is reasonably close to the initial amount of 2.634 pmol of free HSA present in the sample. To investigate the effects of interference by other proteins on the performance of the proposed ZnSe QD-based sensors, we performed a competitive assay for HSA detection in platelet-poor human plasma. The concentration of HSA in human serum plasma of healthy adults is typically between 600 and 800 mM. In order to avoid using large quantities of sensor, the human serum plasma was diluted 1000 times, thus making the expected HSA concentration in the samples to be between 0.6 and 0.8 mM. To perform the assay manually, 12 vials each containing 100 mL of the 1000  diluted human serum plasma were prepared. The vials were dosed with increasing amounts of ZnSe QD-HSA and free antiHSA and the fluorescence emission spectrum of each sample was recorded. The results are plotted in Fig. 6. When the peak fluorescence

1.2 Normalized Peak Emission Intensity

148

1 0.8 0.6 0.4 QD-HSA QD-HSA/anti-HSA Immunoassay

0.2 0

0

0.04 0.08 Concentration of Labeled HSA (µM)

0.12

Fig. 6. Homogeneous competitive immunoassay that detects free HSA in human serum plasma by dosing the sample with progressively increasing amounts of ZnSe QD-HSA sensor and anti-HSA antibody.

emission intensity reached the sensorþantibody calibration line, a total of 30 mL of 1.811 mM ZnSe QD-HSA sensor solution and 65 mL of 1.93 mM anti-HSA solution had been added to the sample. The difference between the anti-HSA and labeled HSA was 0.071 nmol, which corresponds to the amount of free HSA contained in the 100 mL of 1000  diluted human serum plasma. This means that the 1000  diluted plasma has a concentration of HSA equal to 0.71 mM, which is within the expected range for healthy adults. The dynamic range of detection in homogeneous assays using the QD-based sensors described in this paper can be determined by identifying the lowest concentration of sensor that can be unambiguously detected by the instrument being used over the background fluorescence of the sample and the highest concentration of sensor that can be detected before interference between neighboring QDs adversely affects the optical signal from the sample (Mei et al., 2008). The limit of detection for a specific QD-based sensor was obtained by performing a dilution series and measuring the peak fluorescence emission intensity of the sensor as function of concentration of the sensor in solution. For example, ZnSe QDs with quantum yield of  25% before surface modification resulted in a limit of detection equal to 0.5 nM HSA for the ZnSe QD-HSA sensor. The upper limit of detection was determined by the maximum concentration of QDs in solution before the onset of optical interference (Mei et al., 2008). This interference typically appears at a particle volume fraction of the order of 5  10  5 that corresponds to a QD particle number concentration of the order of 1.5  1015 particles/mL (2.5 mM), when the average size of the ZnSe QDs is  4 nm (Mei et al., 2008). The major advantages of the homogeneous assays discussed in this paper are their speed and simplicity. They are typically one to three orders of magnitude less sensitive than ELISA-type assays, which involve separation of the sensorþtarget complex before detection and can achieve a lower limit of detection in the pM range. The proof-of-principle assays discussed in this paper were repeated successfully using different batches of QDs. The concentration of the target in the samples was measured accurately in every case that we tested, despite the expected batch to batch variations in peak emission intensity that may arise during sensor preparation. Finally, the stability of the ZnSe QD-based sensors was investigated by monitoring the peak emission intensity of QD-labeled antigens and antibodies as function of storage time at 4 1C and under dark conditions. The observed peak emission intensity of the QDs slightly increased over a period of 5 days and subsequently decreased very

J. Wang, T.J. Mountziaris / Biosensors and Bioelectronics 41 (2013) 143–149

slowly with time in storage reaching 80% of the initial value after two months in storage.

4. Conclusions Optical biosensors employing ZnSe QDs that are conjugated to biological probe molecules were developed and tested in homogeneous (separation-free) immunoassays. The detection scheme is based on variations of the peak fluorescence emission intensity of the ZnSe QDs upon binding of the QD-based sensors to specific target proteins. Direct homogeneous immunoassays for bFGF and PSA were performed using ZnSe QD-anti-bFGF and ZnSe QD-antiPSA, respectively, as sensors. A competitive homogeneous immunoassay that employed ZnSe QD-HSA sensors and anti-HSA antibodies was used to detect free HSA in PBS and diluted human serum. The observed changes in the fluorescence emission intensity of the QDs were caused by surface-induced electronic perturbations resulting from the binding of the QD-biomolecule sensors to their targets which alter the negative charge distribution around the QDs. The QD-based optical sensors demonstrated in this work can be potentially attractive for applications requiring rapid quantitative detection of biological targets, e.g., in emergency care and point-of-care diagnostics.

Acknowledgment The authors acknowledge financial support by the University of Massachusetts, via the President’s Science and Technology Fund (UMass NanoMedicine Institute) and the CVIP Technology Development Fund, and by the National Science Foundation, via the Center for Hierarchical Manufacturing (NSF-NSEC CMMI-0531171). References Algar, W.R., Krull, U.J., 2008. Analytical and Bioanalytical Chemistry 391 (5), 1609–1618. Alivisatos, A.P., 1996. Science 271 (5251), 933–937.

149

Bakalova, R., Zhelev, Z., Ohba, H., Baba, Y., 2005. Journal of the American Chemical Society 127 (26), 9328–9329. Brus, L., 1991. Applied Physics A 53 (6), 465–474. Cadars, S., Smith, B.J., Epping, J.D., Acharya, S., Belman, N., Golan, Y., Chmelka, B.F., 2009. Physical Review Letters 103 (13), 136802–136806. Chan, W.C.W., Nie, S.M., 1998. Science 281 (5385), 2016–2018. Chen, W., Xu, D.H., Liu, L.Q., Peng, C.F., Zhu, Y.Y., Ma, W., Bian, A., Li, Z., Yuanyuan, Jin, Z.Y., Zhu, S.F., Xu, C.L., Wang, L.B., 2009. Analytical Chemistry 81 (21), 9194–9198. Clapp, A.R., Medintz, I.L., Uyeda, H.T., Fisher, B.R., Goldman, E.R., Bawendi, M.G., Mattoussi, H., 2005. Journal of the American Chemical Society 127 (51), 18212–18221. Gill, R., Zayats, M., Willner, I., 2008. Angewandte Chemie International Edition 47 (40), 7602–7625. Goldman, E.R., Clapp, A.R., Anderson, G.P., Uyeda, H.T., Mauro, J.M., Medintz, I.L., Mattoussi, H., 2004. Analytical Chemistry 76 (3), 684–688. Goldman, E.R., Mattoussi, H., Anderson, G.P., Medintz, I.L., Mauro, J.M., 2005a. Methods in Molecular Biology 303, 19–34. Goldman, E.R., Medintz, I.L., Mattoussi, H., 2006. Analytical and Bioanalytical Chemistry 384 (3), 560–563. Goldman, E.R., Medintz, I.L., Whitley, J.L., Hayhurst, A., Clapp, A.R., Uyeda, H.T., Deschamps, J.R., Lassman, M.E., Mattoussi, H., 2005b. Journal of the American Chemical Society 127 (18), 6744–6751. Hermanson, G.T., 1996. Bioconjugate Techniques, San Diego. Hines, M.A., Guyot-Sionnest, P., 1998. The Journal of Physical Chemistry B 102 (19), 3655–3657. Kerman, K., Endo, T., Tsukamoto, M., Chikae, M., Takamura, Y., Tamiya, E., 2007. Talanta 71 (4), 1494–1499. Kricka, L.J., 1994. Clinical Chemistry 40 (3), 347–357. Ligler, F.S., 2009. Analytical Chemistry 81 (2), 519–526. Medintz, I.L., Clapp, A.R., Mattoussi, H., Goldman, E.R., Fisher, B., Mauro, J.M., 2003. Nature Materials 2 (9), 630–638. Mei, B.C., Wang, J., Qiu, Q., Heckler, T., Petrou, A., Mountziaris, T.J., 2008. Applied Physics Letters 93 (8), 083114–083117. Michalet, X., Pinaud, F.F., Bentolila, L.A., Tsay, J.M., Doose, S., Li, J.J., Sundaresan, G., Wu, A.M., Gambhir, S.S., Weiss, S., 2005. Science 307 (5709), 538–544. Murray, C.B., Kagan, C.R., Bawendi, M.G., 2000. Annual Review of Materials Science 30, 545–610. Sapsford, K.E., Pons, T., Medintz, I.L., Mattoussi, H., 2006. Sensors 6 (8), 925–953. Sikavitsas, V., Nitsche, J.M., Mountziaris, T.J., 2002. Biotechnology Progress 18 (4), 885–897. Smith, A.M., Ruan, G., Rhyner, M.N., Nie, S.M., 2006. Annals of Biomedical Engineering 34 (1), 3–14. Wang, J., Lei, P., Andreadis, S.T., Mountziaris, T.J., 2012. Analytical Letters 45 (2–3), 227–241. Wang, S.P., Mamedova, N., Kotov, N.A., Chen, W., Studer, J., 2002. Nano Letters 2 (8), 817–822. Woodbury, R.L., Varnum, S.M., Zangar, R.C., 2002. Journal of Proteome Research 1 (3), 233–237.