Functionalized gold nanoclusters as fluorescent labels for immunoassays: Application to human serum immunoglobulin E determination

Author’s Accepted Manuscript Functionalized gold nanoclusters as fluorescent labels for immunoassays: Application to Human serum immunoglobulin E determination María Cruz Alonso, Laura Trapiella-Alfonso, José M Costa Fernández, Rosario Pereiro, Alfredo SanzMedel www.elsevier.com/locate/bios

PII: DOI: Reference:

S0956-5663(15)30336-5 http://dx.doi.org/10.1016/j.bios.2015.08.011 BIOS7907

To appear in: Biosensors and Bioelectronic Received date: 3 June 2015 Revised date: 30 July 2015 Accepted date: 8 August 2015 Cite this article as: María Cruz Alonso, Laura Trapiella-Alfonso, José M Costa Fernández, Rosario Pereiro and Alfredo Sanz-Medel, Functionalized gold nanoclusters as fluorescent labels for immunoassays: Application to Human serum immunoglobulin E determination, Biosensors and Bioelectronic, http://dx.doi.org/10.1016/j.bios.2015.08.011 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Functionalized gold nanoclusters as fluorescent labels for immunoassays: application to human serum immunoglobulin E determination

María Cruz Alonso, Laura Trapiella-Alfonso, José M Costa Fernández*, Rosario Pereiro*, Alfredo Sanz-Medel Department of Physical and Analytical Chemistry, Faculty of Chemistry, University of Oviedo, Julián Clavería, 8. 33006 Oviedo, Spain *Authors to whom correspondence should be addressed. [email protected], [email protected]

ABSTRACT A quantitative immunoassay for the determination of immunoglobulin E (IgE) in human serum using gold nanoclusters (AuNCs) as fluorescent label was developed. Water soluble AuNCs were synthesised using lipoic acid and then thoroughly characterized. The obtained AuNCs have a particle size of 2.7 ± 0.1 nm and maximum fluorescence emission at 710 nm. The synthesized AuNCs showed very good stability of the fluorescent signal with light exposure and at neutral and slightly basic media. A covalent bioconjugation of these AuNCs with the desired antibody was carried out by the carbodiimide reaction. After due optimization of such bioconjugation reaction, a molar ratio 1:3 (antibody:AuNCs) was selected. The bioconjugate maintained an intense luminescence emission, slightly red-shifted as compared to the free AuNCs. Two typical immunoassay configurations, competitive and sandwich, were assayed and their performance for IgE determination critically compared. After the different immunoassay steps were accomplished, the fluorescence emission of the bioconjugate was measured. While the sandwich format provided a detection limit (DL) of 10 ng/mL and a linear range between 25 and 565 ng/mL of IgE, the competitive format revealed a DL of 0.2 ng/mL with a linear range between 0.3 and 7.1 ng/mL The applicability of the more sensitive competitive fluorescent immunoassay was assessed by successful analysis of the IgE in human serum and comparison of results with those from a commercial kit. The main advantages of the proposed AuNCs0

based fluorimetric method include a low DL and a simple immunoassay protocol involving few reagents.

KEYWORDS: Gold nanoclusters, fluorescence, immunoglobulin E, competitive immunoassay, human serum 1. INTRODUCTION Current research in the biomedical field requires fast, sensitive, reliable and reproducible detection of many different biomolecules, being fluorimetric detectors a widely used approach in such biomedical studies (Liu et al., 2013). Among the different types of fluorescent labels used today, the most common ones for biomedical applications are organic dyes and semiconductor quantum dots (QDs) -based fluorophores (Zhang and Wang, 2014). However, organic fluorophores can be highly susceptible to photobleaching (Frangioni, 2003) while semiconductor QDs typically contain toxic metal species (e.g. CdSe) which restrains their use. Recent advances in nanotechnology have introduced a new class of fluorescent labels, the so-called, fluorescent “metal nanoclusters”. These nanoclusters are composed of a few to roughly several hundred metal atoms, with sizes between 0.2 and 3 nm (Le Guevel, 2014). This type of nanostructures is characterized by sizes comparable to the Fermi wavelength of electrons and, therefore, they can exhibit molecule-like properties (Guo and Wang, 2011). The result of this small size is that the continuous density of typical states of metals breaks up into discrete energy levels. Consequently, metal nanoclusters offer new and improved optical, electrical and chemical characteristics with respect to other nanoparticles and bulk material. Among the different metal nanoclusters, gold nanoclusters (AuNCs) exhibit strong photoluminescence, large Stokes shifts, emission in the near infrared region, good photostability, biocompatibility and low toxicity, constituting an interesting alternative to more conventional luminescent markers used in the bioanalytical field (Le Guevel, 2014; Shang et al., 2011; Zhang and Wang, 2014). Moreover, due to its reduced size, several AuNCs can be incorporated inside a single antibody; thus, the goal of obtaining a particle with brighter luminescence is successful. While the synthesis of AuNCs have been well studied in the latest years (Aldeek et al., 2013; Xie et al., 2009), the development of quantitative analytical methodologies using fluorescent AuNCs as labels is still in its infancy (C.L. Liu et al., 2011; Wang et al., 2011). AuNCs have been 1

reported as fluorescent direct sensors for several applications (e.g. mercury sensing in living cells (Shang et al., 2012)). However, the most important field of AuNCs applications in bioscience is the development of bioassays based in the conjugation of a selective receptor, e.g., antibody (Ab) or aptamer, to the surface of the nanocluster. Although a few examples have been reported for proteins and nucleic acids imaging (Lin et al., 2009), the applicability of AuNCs as fluorimetric labels for the development of quantitative immunoasays is still scarce (Peng et al., 2012). Immunoglobulin E (IgE) is a glycoprotein involved in allergy processes (type 1 hypersensitivity or immediate hypersensitivity). IgE is the immunoglobulin component found in the lowest levels in human serum, with baseline levels of 0.3 µg/mL (J.M. Liu et al., 2011; Papamichael et al., 2007), but it increases notably its concentration in serum when the organism experiments allergic processes (Hamilton and Adkinson, 2003; Kreuzer et al., 2001). There are several current methods to determine IgE in serum, including radioimmunometric assays or spectrophotometric ELISA. The detection limit (DL) of the radioimmunometric assay is very low (0.05 ng/mL) (Poulsen et al., 1986), but using radioactive reagents constitutes its more serious drawback. The spectrophotometric ELISA, on the other hand, is commonly used in commercial test kits for human serum IgE determination even if its DL (12 ng/mL) is much poorer (Papamichael et al., 2007). Fluorescent methods described so far in the literature for IgE determination are scarce and include the use of a DNA aptamer-based bionalysis approach with final detection by fluorescent polarization (Gokulrangan et al., 2005). This latter method’s DL was 65 ng/mL and no application to real samples was reported. Also, a solid substrate room temperature phosphorescence (RTP) immunoassay using CdSe@CdSQDs-cysteine has been reported more recently (J.M. Liu et al., 2011). Unfortunately, even if the given DL was very low (3.0 × 10−4 ng/mL), the procedure is rather cumbersome (e.g. the addition of Pb2+ as well as the absence of moisture to develop RTP signals are needed). In this work, AuNCs were synthesized via lipoic acid as stabilizing ligand, and then characterized by different physico-chemical complementary techniques. After a proper carbodiimide conjugation (Hermanson, 2008) with the desired antibody, competitive and sandwich immunoassay formats were evaluated. The selected approach was investigated for the advantageous determination of IgE in human serum as “model” analyte. 2

2. EXPERIMENTAL 2.1. Materials AuCl4Na.H2O (99% powder; Sigma-Aldrich, Milwaukee, WIS, USA) as metal precursor, lipoic acid (> 98% powder; Across Organics, Geel, Belgium) as surface ligand and NaBH4 (> 98% powder, Sigma-Aldrich) as reducing agent were employed for the synthesis of gold nanoclusters. The competitive immunoassay format was performed with a polyclonal antihuman IgE antibody ((ε-chain specific) produced in goat; Sigma-Aldrich). The sandwich assay was performed with a monoclonal anti-human IgE antibody (GE-1, produced in mouse; Sigma-Aldrich) for detection and the capture was done with the polyclonal anti-human IgE antibody mentioned above. Human IgE protein (whole molecule, Abcam, Cambridge, UK) was used as standard. 1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) (98% powder; Across Organics) and N-Hydroxysuccinimide (NHS) (95% powder; Thermo Scientific, Waltham, MA, US) were employed for the bioconjugation procedure. For purification steps, Amicon ultra centrifugal filter units (10 and 100 kDa pore size, Millipore, Darmstadt, Germany) were used. Immunoassays were conducted in poly-L-lysine surface coated microscope slides (Electron Microscopy Sciences, USA) with an adhesive on one side press-to-seal silicon isolator (Grace bio-labs, Oregon, USA). After human IgE immobilization (competitive format) or capture Ab (sandwich format), bovine serum albumin (BSA) (99% powder, Merck KgaA, Darmstadt, Germany) was used as blocking agent. Washing steps were carried out with a solution of phosphate buffered saline (PBS) (Sigma-Aldrich). For specificity studies, human serum albumin (HSA), immunoglobulin G from human serum (IgG) and immunoglobulin A from human serum (IgA), all provided by Sigma, were also employed. The functional antigen recognition property of the bioconjugate was measured via a spectrophotometric immunoassay. For this purpose, a polyclonal anti-Goat IgG horseradish peroxidase conjugated Ab (whole molecule, produced in rabbit; SigmaAldrich) was used as a secondary antibody. A TMB substrate kit (Thermo Scientific)

3

was used as peroxidase substrate. This immunoassay was performed in ELISA microtitration plates (96 well; Sterilin Limited, UK). Commercial human serum (from human male AB plasma, USA origin, sterilefiltered, Sigma-Aldrich) was used to investigate the applicability of the immunoassay. To validate the methodology, the IgE (Human) ELISA Kit (Abnova, Taiwan) was used. Four clinical serum samples were obtained from HUCA Hospital (Oviedo, Spain) after duly signed informed consent from the participants. Other chemicals were aqua regia, prepared by mixing concentrated nitric acid (65 %, Merck, Darmstadt, Germany) and hydrochloric acid (37%, Fisher Scientific, MA USA) in a volume ratio of 1:3, used for AuNCs digestion prior to ICP-MS analysis, NaOH (> 97%, Merck) and H2SO4 (Fisher Chemical). Milli-Q deionised ultrapure water, resistivity 18.2 MΩ cm-1 was used throughout.

2.2. Instrumentation Images for the immunoassays were taken with a confocal microscope (DM IRE2; Leica, Germany) with a 63x oil immersion objective. A laser excitation line of 405 nm was used. Imaging was not the goal of our experiments. We resorted to the confocal instrument in order to take advantage of the use of a laser excitation source and we integrated the emission from the whole well surface to obtain the analytical signal. Luminescent measurements in solution were carried out using a fluorescence spectrophotometer (LS-50-B, Perkin Elmer, USA) and the experiments were performed with fluorescence Suprasil® quartz cuvettes (model 101-Qs, Hellma, Mülhein, Germany). Inductively coupled plasma-mass spectrometry (ICP-MS) (Model 7500; Agilent, USA) was used to determine the gold concentration. A high-resolution transmission electron microscope (HR-TEM) (JEOL JEM-2100F; Tokyo, Japan) was employed for morphological analysis of the nanoclusters. This model is equipped with a highresolution CCD camera and an energy x-ray microanalyzer (EDX). The UV-VIS microplate measurements were obtained with an absorbance microplate reader (ELX800; Bio-Tek, USA). An ultrasonic bath (J.P. Selecta, Barcelona, Spain) and a centrifuge (Biofuge Stratus, Heraeus, Germany) were other instrumentation used. Fluorescence data processing and immunoassay curves representation were carried out using Image J software and SOFTmax Pro software (Molecular Devices, Ismaning, Germany), respectively. 4

2.3. Synthesis and Characterization of AuNCs Synthesis of AuNCs was carried out using a slightly modified approach previously described (Aldeek et al., 2013). Briefly, 30 µmol of ligand (lipoic acid) was dissolved in 20 mL of deionized water with 50 µL of 2 M NaOH in a light protected vial. The mixture was then homogenised with the help of an ultrasonic bath for 5 min to facilitate the dissolution of the ligand. Later, 200 µL of 50 mM HAuCl4.3H2O (1:3 final ratio of Au:ligand) was added and immediately the stirring was started. After 5 min, 400 µL of the NaBH4 solution (50 mM) was added to the vial dropwise. The reaction was kept at room temperature with constant stirring for 15 h. Then, a purification step to remove the excess of ligand and reductant was performed by ultrafiltration (10 kDa pore size Amicon ultra centrifugal filter units) with the following cycle sequence: a first cycle at 1600 g for 5 min, a second cycle at 1600 g for 10 min, and finally three washing steps with H2O at 1600 g for 10 min. The purified solution was stored at room temperature protected from light. The final product was first characterised by registering its fluorescence.

2.4. Synthesis of Anti-human IgE/AuNCs Bioconjugate Bioconjugation was carried out in an Eppendorf tube with constant stirring at room temperature where 100 µL of Anti human-IgE antibody (100 µg/mL) was added followed by 283 µL of AuNCs solution (molar ratio of Ab:AuNCs = 1:3). As soon as the AuNCs solution was poured, the mechanical stirring was started. Then, EDC and NHS reagents were added: EDC was in excess, in a molar ratio of Ab:EDC of 1:1500, while the NHS amount was related to the amount of EDC (so that they were in a molar ratio EDC:NHS of 1:1). After 2 h at room temperature with constant stirring, a final purification step was carried out to remove the excess of reagents. This purification was performed by ultrafiltration (100 kDa pore size Amicon ultra centrifugal filter units). The centrifugation sequence consisted of a first cycle at 700 g for 7 min and two washing steps with H2O at 700 g for 7 min. The final purified solution was stored at 4 ºC.

2.5. Competitive Immunoassay Format The optimized procedure for the competitive immunoassay requires in a first step the incubation of the bioconjugate Anti-human IgE polyclonal Ab-AuNCs (75 µL of 10 5

µg/mL) with 75 µL of the standards or samples containing IgE (different IgE concentrations were assayed from 0 µg/mL up to 3 µg/mL) in an Eppendorf tube. Simultaneously, wells on poly-L-lysine coated microscope slide were filled with a solution of IgE (10 µg/mL, 100 µL/well). The IgE immobilization on the microscope slides was achieved by chemical reaction with EDC (adding an excess of EDC in a ratio of 1500 mol of EDC per mol of IgE) and incubation during 2 h at 37 ºC. Later, the solution containing the excess of IgE was removed and the blocking agent solution (1% BSA solution in 10 mM PBS, pH 7.4; 200 µL/well) was added to each well. The microscope slide was then incubated for 2 h at room temperature. Finally, all wells were washed three times with 10 mM PBS pH 7.4 + 0.05% Tween 20 (200 μL/well) to remove the excess of reagents. The mixture of the Eppendorf tubes (sample + bioconjugate) was then added to the wells and incubated for half an hour at 37 ºC. After a washing step with PBS-Tween 20, the fluorescence emission was measured by confocal laser microscopy. Fig. S1 (collected in the Supplementary Information) depicts the schematic representation of the immunoassay procedure.

2.6. Sandwich Immunoassay Format The optimized sandwich-type immunoassay procedure requires a coating of the microscope slide wells with the polyclonal Anti-human IgE antibody (0.5 µg/mL, 100 µL/well) which was incubated for 2 h at 37 ºC. Similarly to the competitive immunoassay, this immobilization was also produced by chemical reaction with EDC, adding an excess of EDC in a ratio of 1500 mol of EDC per mol of IgE. Then, the Ab solution was removed and the blocking solution was added and incubated for 2 h at room temperature. Later, the solution with the analyte (IgE at different concentrations: 0 µg/mL-20 µg/mL; 100 µL/well) was incubated at 37 ºC for half an hour. After a washing step, the immunoprobe (1 µg/mL, 100 µL/well) was added and incubated for half an hour at 37 ºC. After a final washing step to remove the excess of immunoprobe, the fluorescence emission of the bioconjugate was measured by confocal laser microscopy.

2.7. Spectrophotometric Immunoassay

6

Spectrophotometric immunoassays carried out in several studies for the AuNCs fluoroinmunoasays optimization were performed using a secondary Ab labelled with horseradish peroxidase. For the competitive format, the IgE whole molecule was immobilized (10 µg/mL, 100 µL/well) in an ELISA plate and incubated for 6 h at 37 °C. Then, the Ab solution was removed and the blocking agent solution (1% BSA solution in 10 mM PBS, pH 7.4; 200 µL/well) was added. The plate was incubated for 2 h at room temperature. Subsequently, the anti-human IgE polyclonal antibody (10 µg/mL, 100/well) was added and incubated for 2 h at 37 ºC. After a washing step, the anti-goat IgG peroxidaselabelled Ab (dilution 1:1000 in a buffer of PBS-BSA-Tween 20, 100 well) was added and incubated with the same conditions used above. Finally, the plate was washed and the kit TMB substrate was added (TMB/Hydrogen peroxide in a 1:1 ratio, 100 µL/ well). To stop the enzymatic reaction 2 M sulphuric acid (100 µL /well) was added, causing a colour change from blue to yellow. The signal measurement was obtained in an absorbance plate reader (Elix 800; λ = 490 nm). In the sandwich format, the ELISA plate was coated with the Anti-human IgE polyclonal (0.5 µg/mL, 100 µL/well) and blocked under the same conditions and procedure mentioned above. In this case, a new incubation step with the analyte was included (100 µL/well; different concentrations; 0.03 µg/mL-10 µg/mL) for 2 h at 37 ºC. After a washing step, the monoclonal Anti-human IgE antibody (1 µg/mL, 100 µL/well) was added and incubated for 2 h at 37 °C. After a final washing step, the secondary Ab (labelled with the peroxidase) was added and the procedure described above was followed.

2.8. Human Serum Analysis Commercial human serum and four clinical serum samples (P1, P2, P3, P4) were used to assess the applicability of the AuNCs based fluoroimmunoassay offering the best analytical performance. P4 was IgE doped in order to have a sample with a higher concentration. It should be noted here that it was not performed any sample pretreatment, except for a serum dilution with Milli-Q water (necessary to fit the basal levels of the IgE into the calibration curve). To fit into the method linear range the commercial human serum, P1 and P3 samples a 1:1000 serum dilution was required. For P2 the dilution was 1:100 and for P4 it was 1:2000.

7

3. RESULTS AND DISCUSSION 3.1. Synthesis and Characterization of AuNCs The AuNCs were synthesized following the methodology described in the Experimental section. Different photoluminescent excitation/emission spectra were recorded to find those most favourable for AuNCs fluorescence detection. Highest fluorescence signals were observed at 710 nm with excitation at 400 nm (see continuous line of Fig. 1); the emission signal recorded at approx. 800 nm corresponds to the second order scattering of excitation wavelength. The quantum yield (QY) of the synthesized AuNCs was evaluated according to a relative method based on the standard rhodamine 6G (see Supporting Information) and a value of 5.2% was obtained. For application of the synthesised AuNCs as labels in quantitative bioanalysis it is of great value to determine the nanoparticle concentration (e.g. number of AuNCs in a given volume of solution). For such purpose, the total mass of Au in the sample containing the AuNCs must be first determined. Thus, the amount of gold present in the AuNCs suspensions, after the synthesis process, was determined by ICP-MS in three different solutions: a stock solution which contained the synthesis precursor (AuCl4Na.H2O), the AuNCs purified solution and the filtered solution collected during the purification process. A mass balance (expressed in moles) was calculated from the results to obtain the synthesis molar yield, which turned out to be 72 ± 3%. In order to reliably optimize the bioconjugation process it is necessary to know the particle concentration of the AuNCs solution (e.g. number of AuNCs in a given volume). For such purpose, the AuNCs were characterized by HR-TEM. They presented a narrow size distribution (see Fig. 2A). A crystalline structure of the material, face-centred cubic structure of gold atoms, was observed (Fig. 2B) by the pattern of selected area diffraction (SAED). EDX measurements showed that the major elements were gold, oxygen and sulphur (the presence of sulphur should correspond to lipoic acid, the ligand employed to stabilise the AuNCs). Using the images obtained by HR-TEM, the average size of the AuNCs could be calculated. For this purpose, measurements of nanocluster diameters were randomly taken from the captured pictures. 200 individual values were measured, revealing an average size of 2.7 ± 0.1 nm (which corresponds to a AuNC volume of 9.81 nm3, assuming an spherical shape). On the other hand, as the crystal structure was face-centred cubic, it was possible to calculate the volume occupied by a unit cell with such structure, by taking into account the density of gold (19.3 g/cm3) and the mass of a unit cell (1.31x10-21 g/cell). With 8

these reference values, the unit cell volume obtained was 6.79 × 10−2

"#3 $%&&

. By dividing

the AuNC volume by the unit cell volume it was obtained an average value of 145 crystal lattices/NC. As a face-centred cubic structure contains 4 gold net atoms per unit cell, it was determined that each nanocluster is composed by 579 atoms of gold. These experimental results are in agreement with the size value given in the literature (Jin, 2010; Li et al., 2015). Combining the gold molar concentration in the AuNCs suspension obtained by ICP-MS (0.41 mM) and the results mentioned above from the HR-TEM, the synthesized AuNCs particle concentration present in the purified solution was calculated to be 4.26 ∗ 10()

*+, -/

. This result is an important parameter because it allows

to determine the molar ratio to perform the conjugation of the AuNCs with a biomolecule (in this case, an antibody). Studies were conducted to evaluate the AuNCs stability. Fluorescence intensity of the synthesized AuNCs remained constant at least over seven days of ambient light exposure. Moreover, we have observed that no significant fluorescence signal changes are observed for at least six months under darkness. The fluorescence intensity of AuNCs can be influenced by the pH of the medium (Richards et al., 2008; Shiraishi et al., 2002) and our results on fluorescence emission at different pHs showed highest fluorescence intensities for neutral-slightly basic media (see Fig. S2 in Supplementary Information).

3.2. Synthesis of the Anti-human IgE-AuNCs bioconjugate The process of bioconjugation of the AuNCs (acting as fluorescent labels) to the adequate antibodies was optimized allowing to maintain the expected properties (photoluminescence and molecular recognition, respectively) of both components. In this vein, several Ab:AuNC molar ratios (1:1, 1:3, 1:5, 1:7 and 1:10) were assayed. Fig. 3 shows that signals obtained for 1:10 ratio is lower than for 1:7. This can be attributed to the fact that a high AuNCs concentration in the reaction medium could induce aggregation of the particles and consequently fewer AuNCs are available for bioconjugation. According to Fig. 3, the best fluorescent signals occurred at 1:3, 1:5 and 1:7 molar ratios. A more detailed study, with the selected three molar ratios (1:3, 1:5 and 1:7) was made by spectrophotometry in order to check the functionality of the antibody after the bioconjugation with AuNCs, in each case. In addition, the study was 9

also performed with unlabelled Ab (at the same concentration as in the bioconjugate). The observed results showed lower absorbance signals for those bioconjugates with higher ratios of AuNCs, which could be attributed to the fact that high content of AuNCs may block the recognition sites of the antibody. A molar ratio of 1:3 was eventually chosen as labelling conditions for the development of the IgE fluoroimmunoassay. Finally, the bioconjugate was characterized by fluorescence, obtaining the maximum emission signal at 740 nm (see dotted line in Fig. 1). Similar fluorescence emission shifts of the AuNCs fluorescence emission, after bioconjugation to antibodies, have been also observed for other nanomaterials after binding to proteins (Sapsford et al., 2011).

3.3. IgE Determinations Based on a Competitive Format In this assay format, the analyte (IgE) in the sample and the previously immobilized IgE compete for the binding sites of the immunoprobe (Anti-human IgE antibody-AuNCs). Thus, as the IgE concentration in the sample increases, the concentration of free bioconjugate capable of interacting with the immobilized IgE decreases and, so, the fluorescence signal was lower (producing a curve of the inverse sigmoid type). Two studies were conducted to optimize the immobilized IgE and the Ab-AuNCs bioconjugate concentrations following the same immunoassay protocol described in the Experimental section (without adding the standard that contains IgE in order to get the maximum fluorescence signal): in the first study, the amount of Ab-AuNCs/well was kept constant at 2.5 pmol (75 mL of 5 µg/mL per well) and the amount of immobilized IgE was varied from 0.52 pmol (100 mL of 1 µg/mL Ig E per well) up to 10.5 pmol (100 mL of 20 µg/mL IgE per well). With this set of conditions, the maximum fluorescence was obtained at a 1:1 mol ratio of immobilized IgE:bioconjugate (Fig. 4A). In the second study, the moles of bioconjugate and of IgE immobilized were increased simultaneously (keeping a 1:1 mol ratio). It was found that for 5.3 pmol and higher, the fluorescence intensity stayed constant (Fig. 4B). Therefore, the optimum values selected for this format were: 5.4 pmol of bioconjugate and 5.3 pmol of immobilized IgE. The coating procedure to immobilize IgE on the plate was also studied: adsorption coating (keeping the incubation at 37 °C for 6 h) and reaction with EDC (37 °C, 2 h). In

10

both cases, a blank was made (i.e. without adding the label). From the microscopy images, it was observed that no fluorescence occurred without IgE. Also, improved results were obtained with a coating by reaction with EDC (in comparison to passive adsorption). The analytical performance characteristics were assessed by the analysis of a series of IgE standard solutions at different concentration levels, using the optimized conditions. The corresponding inhibition curve was fitted using a four-parameter equation with SOFTmax Pro software. The principal parameters are defined by different inhibitory concentration (IC) values calculated from the inhibition curve. The detection limit (IC10) was 0.2 ng/mL of IgE (15 pg per well) and the linear range (IC20-IC80) extended from 0.3 up to 7.1 ng/mL of IgE.

3.4. IgE Based on a Sandwich Format In the sandwich format the polyclonal IgE antibody was used as the capture Ab, while the monoclonal Ab was employed for recognition and final detection. As customary, the concentration of immobilized polyclonal Anti-Human IgE antibody and the concentration of Ab-AuNCs bioconjugate were optimized, using an established spectrophotometric immunoassay in order to observe interactions between the two Anti Human-IgE antibodies. The initial concentrations assayed were 3.3 pmol (100 mL of 5 µg/mL) of Anti-Human IgE polyclonal antibody and 5 pmol (100 mL of 7.5 µg/mL) of Anti-IgE monoclonal antibody. Three analyte concentrations were studied: no added analyte (0 µg/mL), a low concentration of 10-3 µg/mL (5.41 x 10-3 nM) and a higher concentration of 1 µg/mL (5.41 nM) of IgE. In order to check the blocking solution efficiency the “control” had no immobilized polyclonal. The results showed that the blocking solution worked perfectly, because there was not significant signal for the “control”. However, when IgE was not added it was observed a high absorbance signal, which means that the two anti-Human-IgE antibodies recognized each other. By lowering the concentrations of the two antibodies (capture and recognition) the interaction between them was reduced. Quantities eventually selected were 0.33 pmol (100 mL of 0.5 µg/mL) of Anti-Human IgE polyclonal Ab and 0.66 pmol (100 mL of 1 µg/mL) of Anti Human IgE monoclonal-AuNCs bioconjugate. The corresponding calibration curve was obtained for such optimized conditions (in this sandwich format, of course, the calibration curve shape was the inverse of the curve

11

observed in the competitive format) and it is plotted in Fig. S3 of Supplementary Material. For this format the obtained DL was 10 ng/mL and the linear range extended from 25 up to 565 ng/mL of IgE.

3.5. Application to Analysis of IgE in Human Serum The applicability of the more sensitive and simpler competitive AuNCs-based fluorescent immunoassay was assessed for the determination of IgE in a commercial human serum sample and four clinical serum samples. First, in order to assess the influence of other possible coexistent species in the real samples, a specificity study was done. The interfering proteins that were tested include two other immunoglobulins (human IgG and human IgA) and an abundant protein in human serum (HSA). Moreover, it was investigated the effect of BSA (a protein used along the immunoassay protocol). For the study, samples containing IgE at a concentration level of 0.3 mg/mL alone or in the presence of the tested potential interferences at common concentration levels present in human serum were analysed (also a BSA concentration similar to HSA was tested). Results are given in Fig. S4 of Supplementary Information and show that the analytical signal for the IgE in the presence of interferences differs by less than 5% from the signal of the standard containing only IgE, thus certifying the high selectivity of the method. A 1:1000 dilution of the commercial serum was required to ensure sample IgE concentration levels within the linear range of such format. Known IgE concentrations were added to this diluted sample. Fig. 5 collects the competitive immunoassay curves obtained using IgE standards at different concentrations prepared in PBS solution. The same Figure shows data obtained when IgE, at different concentrations in diluted human serum, was present. As can be seen, both fluorescent curves overlap. Therefore, it is possible to quantify IgE in an unknown diluted serum sample when using a calibration curve made in PBS. A recovery of 78.2 ± 4.6% was obtained following Eq. 1: [?@A]BCDEFG[?@A]HIJEK M∗ [?@A]JFFLF

5%$8:%;< = >

100

(1)

Table 1 collects the IgE concentration finally obtained in the undiluted analysed commercial serum with the fluorescent AuNCs immunoassayed proposed as well as those achieved for the four assayed clinical samples. For validation of the new AuNCsbased fluorescent methodology, the determination of IgE in the same human serum samples was also performed with a commercial IgE Human ELISA kit (see Table 1). 12

The IgE concentrations obtained were statistically indistinguishable from the value obtained using the here proposed fluorescent immunoassay. These results, together with the higher sensitivity, the fewer steps and reagents required (as compared with current commercial kits), indicates towards a high potential of the developed AuNCs-based competitive format for the determination of IgE in human serum. 4. CONCLUSIONS Characterization of AuNCs synthesized in our laboratory has been carried out by combining fluorescence spectrometry, HR-TEM and ICP-MS techniques. Synthesis yield, size, size distribution and AuNCs particle concentration figures were obtained. This information is crucial to develop reproducible AuNCs bioconjugation methods. In order to achieve an optimum Ab:AuNC molar ratio allowing Ab functionality and keeping the luminescent properties of the AuNCs, the bioconjugation procedure was optimized through fluorescence measurements and also with a spectrophotometric immunoassay. The comparison of analytical parameters derived from calibration curves for the two immunoassays configurations developed (competitive and sandwich) revealed a better DL for the competitive format (0.2 and 10 ng/mL, respectively), being in both cases the detection limits better than those obtained with current commercial ELISA kits (12 ng/mL) (Papamichael et al., 2007). The applicability of this new AuNC immunoprobe was successfully demonstrated for the determination of IgE in human serum, using the competitive configuration. Moreover, the developed competitive format offers other interesting advantages such as a fewer reagents involved and simple procedures. The results here obtained with fluorescent AuNCs as labels point to a promising approach for many novel applications in the field of protein determination using immunoassays.

Acknowledgements This work was supported by project FC-15-GRUPIN14-092 (Principado de Asturias) and MINECO-13-CTQ2013-49032-C2-1-R (Ministerio de Educación y Ciencia). The authors show their gratitude to “The Biotechnology Preparative” and “Photon Microscopy and Image Processing” units of the Scientific-Technical Services of the University of Oviedo for services. 13

References Aldeek, F., Muhammed, M.A.H., Palui, G., Zhan, N., Mattoussi, H., 2013. ACS Nano 7, 2509–2521. Frangioni, J., 2003. Curr. Opin. Chem. Biol. 7, 626–634. Gokulrangan, G., Unruh, J.R., Holub, D.F., Ingram, B., Johnson, C.K., Wilson, G.S., 2005. Anal. Chem. 77, 1963–1970. Guo, S., Wang, E., 2011. Nano Today 6, 240–264. Hamilton, R.G., Adkinson, N.F., 2003. J. Allergy Clin. Immunol. 111, S687–S701. Hermanson, G.T., 2008. The Chemistry of Reactive Groups, in: 2nd Edition, Bioconjugate Techniques. Academic Press pp. 169–212. Jin, R., 2010. Nanoscale 2, 343–362. Kreuzer, M.P., O’Sullivan, C.K., Pravda, M., Guilbault, G.G., 2001. Anal. Chim. Acta 442, 45–53. Le Guevel, X., 2014. IEEE J. Sel. Top. Quantum Electron. 20, 6801312. Li, H., Li, L., Pedersen, A., Gao, Y., Khetrapal, N., Zeng, X.C., 2015. Nano Lett. 15, 682–688. Lin, C.A.J., Yang, T.Y., Lee, C.H., Huang, S.H., Sperling, R.A., Zanella, M., Li, J.K., Shen, J.L., Wang, H.H., Yeh, H.I., Parak, W.J., Chang, W.H., 2009. ACS Nano 3, 395–401. Liu, C.L., Wu, H.T., Hsiao, Y.H., Lai, C.W., Shih, C.W., Peng, Y.K., Tang, K.C., Chang, H.W., Chien, Y.C., Hsiao, J.K., Cheng, J.T., Chou, P.T., 2011. Angew. Chem. Int. Ed. Engl. 50, 7056–7060. Liu, J., Liu, C., He, W., 2013. Curr. Org. Chem. 17, 564–579. Liu, J.-M., Lin, L., Liu, Z.B., Yang, M.L., Wang, X., Zhang, L., Cui, M., Jiao, L., 2011. J. Fluoresc. 22, 419–429. Papamichael, K., Kreuzer, M., Guilbault, G., 2007.. Sensors Actuators B Chem. 121, 178–186. Peng, J., Feng, L.N., Zhang, K., Li, X.H., Jiang, L.P., Zhu, J.J., 2012. Chem. A Eur. J. 18, 5261–5268. Poulsen, L.K., Malling, H.J., Søndergaard, I., Weeke, B., 1986. J. Immunol. Methods 92, 131–136. 14

Richards, C.I., Choi, S., Hsiang, J., Antoku, Y., Vosch, T., Bongiorno, A., Tzeng, Y., Dickson, R.M., 2008. J. Am. Chem. Soc. 130, 5038–5039. Sapsford, K.E., Tyner, K.M., Dair, B.J., Deschamps, J.R., Medintz, I.L., 2011. Anal. Chem. 83, 4453–88. Shang, L., Dong, S., Nienhaus, G.U., 2011. Nano Today 6, 401–418. Shang, L., Yang, L., Stockmar, F., Popescu, R., Trouillet, V., Bruns, M., Gerthsen, D., Nienhaus, G.U., 2012. Nanoscale 4, 4155-4160. Shiraishi, Y., Arakawa, D., Toshima, N., 2002. Eur. Phys. J. E. Soft Matter 8, 377–383. Wang, Y., Chen, J., Irudayaraj, J., 2011. ACS Nano 5, 9718–9725. Zhang, L., Wang, E., 2014. Nano Today 9, 132–157.

FIGURE CAPTIONS Fig.1. Excitation (lem=710 nm for AuNCs and lem=730 nm for bioconjugated AuNCs) and fluorescence emission spectra (λex=400 nm) of AuNCs and bioconjugated AuNCs. In both spectra continuous lines correspond to AuNCs and dotted lines to bioconjugated AuNCs. Fig 2. Results from the morphological study of purified AuNCs by HRTEM. A) Images showing shape and size of AuNCs. B) SAED pattern image to obtain the structure type (face-centred cubic structure). Fig. 3. Influence of Ab:AuNCs molar ratio on measured fluorescence intensity. Fig. 4. Optimization of the competitive immunoassay. A) Study of optimum molar ratio between the immobilized IgE and the bioconjugate (2.5 pmol of AbAuNCs was kept constant for this study). B) Optimization of immobilized IgE and bioconjugate quantities (the number of moles of bioconjugate and of IgE immobilized were increased simultaneously). Fig. 5. IgE calibration graphs, following the competitive configuration, for PBS solution and for serum spiked with different concentrations of IgE (serum dilution 1:1000).

15

Table Table 1. Determination of IgE in a commercial serum sample and in four clinical samples by the proposed AuNCs competitive immunoassay procedure. Results are compared with those achieved with the commercial ELISA kit. Sample

Commercial serum Clinical serum P1 Clinical serum P2 Clinical serum P3 Clinical serum P4

IgE by AuNCs competitive immunoassay (µg/mL) 0.28 ± 0.02 0.44 ± 0.03 0.024 ± 0.005 0.24 ± 0.03 1.36 ± 0.05

IgE by commercial ELISA kit (µg/mL) 0.26 ± 0.01 0.43 ± 0.02 0.027 ± 0.004 0.22 ± 0.03 1.33 ± 0.01

Highlights Water soluble fluorescent AuNCs were synthesized and thoroughly characterized. Bioconjugation of the AuNCs with the corresponding antibody allowed the development of fluorescent immunoassays for sensitive IgE determination. Ø The applicability of the new immunoprobe was demonstrated for the determination of IgE in human serum. Ø The advantageous use of AuNCs as fluorimetric labels for the development quantitative immunoassays has been demonstrated using IgE as model analyte.

Ø Ø

16

Figure(s)

Fig. 1

Fig. 2

B)

A)

Normalized Intensity

Fig. 3

0

20

40

60

80

100

0.8

4.5

6.6

1

7.4 pH

8.6

11.4

B)

A)

Fig. 4

0

20

40

60

80

100

0

20

40

60

80

100

Normalized Intensity

Normalized Intensity

2.6

1 0.2:1

3 2.1:1

13.1

Number of mol of immobilized IgE (pmol)

5.3

Mol ratio of immobilized IgE: Bioconjugate

2 1:1

26.1

4 4.2:1

0 0.00001

20

40

60

80

100

Fig.5

Normalized Intensity

0.0001

0.01

0.1

IgE Concentration (µg/mL)

0.001

1

Serum (1:1000)

PBS

10