Multiplexible Wash-Free Immunoassay Using Colloidal Assemblies of Magnetic and Photoluminescent Nanoparticles Dokyoon Kim,† Hyek Jin Kwon,†,‡ Kwangsoo Shin,†,‡ Jaehyup Kim,§ Roh-Eul Yoo,∥ Seung Hong Choi,†,∥ Min Soh,†,‡ Taegyu Kang,†,‡ Sang Ihn Han,†,‡ and Taeghwan Hyeon*,†,‡ †
Center for Nanoparticle Research, Institute for Basic Science (IBS), Seoul 08826, Republic of Korea School of Chemical and Biological Engineering, Seoul National University, Seoul 08826, Republic of Korea § Department of Physiology, University of Texas Southwestern Medical Center, Dallas, Texas 75390, United States ∥ Department of Radiology, Seoul National University College of Medicine, Seoul 03080, Republic of Korea ‡
S Supporting Information *
ABSTRACT: Colloidal assemblies of nanoparticles possess both the intrinsic and collective properties of their constituent nanoparticles, which are useful in applications where ordinary nanoparticles are not well suited. Here, we report an immunoassay technique based on colloidal nanoparticle assemblies made of iron oxide nanoparticles (magnetic substrate) and manganese-doped zinc sulfide (ZnS:Mn) nanoparticles (photoluminescent substrate), both of which are functionalized with antibodies to capture target proteins in a sandwich assay format. After magnetic isolation of the iron oxide nanoparticle assemblies and their bound ZnS:Mn nanoparticle assemblies (MZSNAs), photoluminescence of the remaining MZSNAs is measured for the protein quantification, eliminating the need for washing steps and signal amplification. Using human C-reactive protein as a model biomarker, we achieve a detection limit of as low as 0.7 pg/mL, which is more than 1 order of magnitude lower than that of enzyme-linked immunosorbent assay (9.1 pg/mL) performed using the same pair of antibodies, while using only one-tenth of the antibodies. We also confirm the potential for multiplex detection by using two different types of photoluminescent colloidal nanoparticle assemblies simultaneously. KEYWORDS: nanoparticles, colloidal assembly, immunoassay, multiplexed, wash-free
P
sensitivity and/or versatility compared with traditional methods.1−26 However, various technical difficulties and practical limitations such as suboptimal reproducibility, nonuniform signal over substrate, lack of multiplex detection capability, and need for custom-designed instruments, have been considered as major challenges that prevent widespread application of nanomaterial-based methods despite their high sensitivity and specificity. We sought to overcome these limitations by adopting nanoparticle assemblies, which is an emerging class of materials with physical and chemical characteristics that arise from both the intrinsic properties of constituent nanoparticles and their interparticle interactions.27−35 They are obtained by assembling nanoparticles in a controlled manner, and their optical, electric, magnetic, mechanical, and biological properties have been
roteins are a useful indicator reflecting the biological states of organisms in that they perform numerous functions, such as regulation of physiological activities, transport of molecules to their respective destinations, and enzymatically catalyzing chemical reactions.1,2 Therefore, accurate assessment of proteins of interest is very important for biomedical research and disease diagnosis.3,4 Immunoassays that rely on the specific interactions between antibodies and antigens are one of the most widely used methods for the quantification of proteins in biological samples. In many popular immunoassay formats, such as enzyme-linked immunosorbent assay (ELISA) or lateral flow assay, target proteins are bound to solid substrate, and a colorimetric or fluorometric signal is generated depending on the concentration of the target proteins.5−7 While these strategies have proven their utility through the successful release of many commercial products, improved methods with increased analytical performance, costeffectiveness, and robustness are still highly desired. Recent advances in nanotechnology have led to the development of immunoassay techniques with greater © 2017 American Chemical Society
Received: June 12, 2017 Accepted: August 8, 2017 Published: August 8, 2017 8448
DOI: 10.1021/acsnano.7b04088 ACS Nano 2017, 11, 8448−8455
Article
www.acsnano.org
Article
ACS Nano
Figure 1. (a) TEM image of 18 nm-sized iron oxide nanoparticles. The inset is a high-resolution TEM image. (b) Magnetic hysteresis loop of the iron oxide nanoparticles measured at 300 K. (c) TEM image of 4.8 nm-sized manganese-doped zinc sulfide (ZnS:Mn) nanoparticles. The inset is a high-resolution TEM image. (d) Normalized one-photon PLE (blue) and PL (red) spectra of the ZnS:Mn nanoparticles.
Figure 2. (a) TEM image of IONAs. (b) TEM image of MZSNAs.
studied extensively for their potential applications including solution-based nanoparticle device fabrication, bioimaging, and drug delivery.31−38 When the size of nanoparticle assemblies is limited within a mesoscopic range (i.e., ∼100 nm to ∼10 μm), they retain the colloidal nature of particles while exhibiting the collective properties of their building blocks,36−50 which provides possibilities in solution-based applications as these colloidal nanoparticle assemblies can replace nanoparticles by virtue of their superior properties. For example, enhanced magnetic resonance imaging and fluorescence imaging have been demonstrated using the colloidal assemblies of nanoparticles.47−50 By taking advantage of such properties, we herein report a wash-free sandwich immunoassay technique using iron oxide nanoparticle assemblies (IONAs) and manganese-doped zinc sulfide (ZnS:Mn) nanoparticle assemblies (MZSNAs) for
protein capture, separation, and quantification. The method features high sensitivity and specificity and also has a potential for multiplex detection. Moreover, thanks to the use of the colloidal nanoparticle assemblies, tunable dynamic range and wash-free procedures without signal amplification steps can be achieved.
RESULTS AND DISCUSSION Iron oxide nanoparticles were synthesized by thermal decomposition of iron oleate. Transmission electron microscopy (TEM) images show the spherical 18 nm-sized iron oxide nanoparticles with a narrow size distribution (Figure 1a and Supporting Information Figure S1). Magnetic behavior of the iron oxide nanoparticles measured at 300 K does not show any hysteresis loop, indicating the superparamagnetic nature of the iron oxide nanoparticles at room temperature (Figure 1b). 8449
DOI: 10.1021/acsnano.7b04088 ACS Nano 2017, 11, 8448−8455
Article
ACS Nano
Figure 3. (a) Schematic illustration of the colloidal nanoparticle assembly immunoassay. Sample solution is added to the solution containing IONAs and MZSNAs. Target protein is bound in between the IONAs and MZSNAs. IONAs are magnetically isolated together with the target protein-bound MZSNAs, and the PL intensity of the solution is measured. (b) Demonstration of the colloidal nanoparticle assembly immunoassay. Before UV irradiation and magnetic isolation, microtubes with or without target protein (CRP) do not show any visual difference (top). Before magnetic isolation, tubes on both sides still do not show any difference under UV irradiation (middle). After magnetic isolation of the colloidal nanoparticle assemblies, only the microtube containing CRP shows the reduced PL under UV irradiation (bottom).
the IONAs except that ZnS:Mn nanoparticles were used instead of the iron oxide nanoparticles. They are spherically shaped with an average size of ∼320 nm and lack the superlattice structure observed in the IONAs (Figure 2b and Supporting Information Figure S7), possibly due to the nonspherical shape and less uniform size of the ZnS:Mn nanoparticles. It is roughly estimated that there are more than 100,000 ZnS:Mn nanoparticles in 1 MZSNA particle (Supporting Information Section S6). Hydrodynamic diameter measured by DLS is ∼390 nm (Supporting Information Figure S7). While the MZSNAs do not show any red shift in their PL spectra due to their large Stokes shift, their PL QY decreased to 9% (Supporting Information Figure S7). In spite of the reduced PL QY, an MZSNA particle is still much brighter than an individual ZnS:Mn nanoparticle due to the number of nanoparticles present. The PAA coating on the surface of the IONAs and MZSNAs provides carboxylic acid groups useful for further functionalization. Biotinylated antibodies were immobilized on the IONAs through the high-affinity streptavidin− biotin interaction,52 where IONAs were functionalized beforehand with streptavidin via 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide (EDC) and N-hydroxysuccinimide (NHS) coupling reaction. Nonbiotinylated antibodies were attached directly onto the MZSNA surface using the EDC/ NHS coupling. The antibodies conjugated on the IONAs and MZSNAs, respectively, were selected from commercially available antibody pairs for sandwich immunoassay. The conjugation results in the increase of the hydrodynamic diameters, and Bradford assay data also supports the conjugation of antibodies on the nanoparticle assemblies (Supporting Information Figure S8). Figure 3a shows a schematic illustration of the wash-free immunoassay using the colloidal nanoparticle assemblies. Briefly, a sample solution is added to a mixture solution of antibody-conjugated IONAs and MZSNAs. Since the antibodies on the IONAs and those on the MZSNAs are selected to bind to different epitopes on the target protein, sandwich structures are formed where the target proteins are positioned in between the IONAs and MZSNAs. After target proteinbound MZSNAs are isolated together with the IONAs by magnetic separation,53,54 the PL intensity of the remaining MZSNAs is measured and used for quantitative analysis. Figure 3b shows a demonstration of the concept, where relatively large amounts of target protein and colloidal nanoparticle assemblies are used to distinguish the PL intensity change with naked eye. The greater brightness of an MZSNA particle than that of a
ZnS:Mn nanoparticles were obtained by reacting zinc chloride and manganese chloride with elemental sulfur in dibenzylamine, leading to near-spherical particles with an average size of 4.8 nm (Figure 1c and Supporting Information Figure S2). Photoluminescence excitation (PLE) and photoluminescence (PL) spectra of the ZnS:Mn nanoparticles were measured (Figure 1d). The strong emission peak at 575 nm, which originates from the 4T1 to 6A1 transition of Mn2+ ions,51 proves the synthesis of high-quality ZnS:Mn nanoparticles. The Mn doping level as determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES) and energy dispersive X-ray spectroscopy (EDS) was around 1% (Supporting Information Figure S2), and the PL quantum yield (QY) of the ZnS:Mn nanoparticles was estimated to be 40%. IONAs and MZSNAs were produced from the self-assembly of the iron oxide and ZnS:Mn nanoparticles, respectively, using an oil-in-water microemulsion system. Briefly, nanoparticles dispersed in chloroform were mixed with deionized water in the presence of dodecyltrimethylammonium bromide (DTAB), and the nanoparticles self-assembled to form the colloidal assemblies following the evaporation of chloroform (Supporting Information Figure S3). The positively charged assynthesized colloidal nanoparticle assemblies were further coated with poly(acrylic acid) (PAA), a negatively charged polyelectrolyte, by electrostatic interactions. Zeta potential measurements reveal that the initial positive charge is changed to negative after the PAA coating (Supporting Information Figure S4). Figure 2a shows a representative TEM image of the IONAs exhibiting polyhedral shapes with an average size of ∼190 nm (Supporting Information Figure S5). The iron oxide nanoparticles act like artificial atoms in the IONAs, forming a highly ordered superlattice structure. Assuming the close packing of the iron oxide nanoparticles, it is calculated that one IONA particle contains ∼700 iron oxide nanoparticles on average (Supporting Information Section S6). Dynamic light scattering (DLS) analysis shows the hydrodynamic diameter of ∼200 nm which is slightly larger than the size measured by TEM (Supporting Information Figure S5). The IONAs also exhibit the superparamagnetic behavior at room temperature (Supporting Information Figure S5). Since an IONA particle has a much larger volume compared with an individual iron oxide nanoparticle, much greater magnetic attraction force is generated under external magnetic field and facilitates efficient magnetic separation. Furthermore, larger surface area of an IONA particle is also advantageous for applications involving surface reactions. The MZSNAs were synthesized similarly to 8450
DOI: 10.1021/acsnano.7b04088 ACS Nano 2017, 11, 8448−8455
Article
ACS Nano
Figure 4. (a) Normalized plot showing the PL versus CRP concentration obtained using MZSNAs and IONAs. (b) Fitted curve (blue line) showing the detection limit of the immunoassay. (c) Plots showing the PL versus CRP concentration obtained using different amounts of the colloidal nanoparticle assemblies (160 ng [Zn] of MZSNAs and 160 ng [Fe] of IONAs were used for blue dots, and 1280 ng [Zn] of MZSNAs and 1280 ng [Fe] of IONAs were used for red dots). Error bars are ±1 standard deviation. (d) Correlation plot showing the CRP concentrations in human serum samples measured using ELISA (y-axis) and the colloidal nanoparticle assembly immunoassay (x-axis) (n = 10 each for healthy controls (blue dots) and inflammatory patients (red dots), R2 = 0.98).
amount of MZSNAs below a certain threshold. We could achieve a detection limit of 0.7 pg/mL (Figure 4b), which is lower than that of ELISA (9.1 pg/mL) performed using the same pair of antibodies (Supporting Information Figure S11), while using only about one-tenth of antibodies (Supporting Information Section S12). The detection limit is defined as a concentration where PL intensity is equivalent to 3 times the standard deviation below the PL intensity of zero-analyte control (blank sample). Another characteristic of the colloidal nanoparticle assembly immunoassay is a tunable dynamic range by adjusting the amount of IONAs and MZSNAs. For example, by increasing the amounts of IONAs and MZSNAs 8 times, a standard curve covering 8 times higher CRP concentrations is obtained (Figure 4c). The colloidal nanoparticle assembly immunoassay shows a good correlation with ELISA for both human serum (collected from inflammatory patients and healthy controls) and CRP-spiked PBS samples (Figure 4d and Supporting Information Figure S13, respectively), and inter- and intra-assay coefficients of variation were
