Article Cite This: ACS Sens. 2017, 2, 1627-1636
pubs.acs.org/acssensors
Tuning the Gold Nanoparticle Colorimetric Assay by Nanoparticle Size, Concentration, and Size Combinations for Oligonucleotide Detection Varsha Sanjay Godakhindi,†,# Peiyuan Kang,‡,# Maud Serre,§,# Naga Aravind Revuru,‡ Jesse Minghao Zou,‡ Michael R. Roner,∥ Ruth Levitz,⊥ Jeffrey S. Kahn,⊥ Jaona Randrianalisoa,¶ and Zhenpeng Qin*,†,‡,▲,○
ACS Sens. 2017.2:1627-1636. Downloaded from pubs.acs.org by IOWA STATE UNIV on 01/17/19. For personal use only.
†
Department of Bioengineering and ‡Department of Mechanical Engineering, The University of Texas at Dallas, Richardson, Texas 75080, United States § Ecole Nationale Supérieure d’Ingénieurs de Reims (ESIReims), University of Reims Champagne - Ardenne, 3 Espl. Roland Garros, 51100 Reims, France ∥ Department of Biology, University of Texas at Arlington, 701 South Nedderman Drive, Arlington, Texas 76019, United States ⊥ Departments of Pediatrics & Microbiology and ▲Department of Surgery, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, Texas 75390, United States ○ Children’s Medical Center, 1935 Medical District Drive, Dallas, Texas 75235, United States ¶ Groupe de Recherche en Sciences pour l’Ingénieur (GRESPI) - EA 4694, University of Reims Champagne - Ardenne, 51687 Reims Cedex 2, France S Supporting Information *
ABSTRACT: Gold nanoparticle (GNP)-based aggregation assay is simple, fast, and employs a colorimetric detection method. Although previous studies have reported using GNP-based colorimetric assay to detect biological and chemical targets, a mechanistic and quantitative understanding of the assay and effects of GNP parameters on the assay performance is lacking. In this work, we investigated this important aspect of the GNP aggregation assay including effects of GNP concentration and size on the assay performance to detect malarial DNA. Our findings lead us to propose three major competing factors that determine the final assay performance including the nanoparticle aggregation rate, plasmonic coupling strength, and background signal. First, increasing nanoparticle size reduces the Brownian motion and thus aggregation rate, but significantly increases plasmonic coupling strength. We found that larger GNP leads to stronger signal and improved limit of detection (LOD), suggesting a dominating effect of plasmonic coupling strength. Second, higher nanoparticle concentration increases the probability of nanoparticle interactions and thus aggregation rate, but also increases the background extinction signal. We observed that higher GNP concentration leads to stronger signal at high target concentrations due to higher aggregation rate. However, the fact the optimal LOD was found at intermediate GNP concentrations suggests a balance of two competing mechanisms between aggregation rate and signal/background ratio. In summary, our work provides new guidelines to design GNP aggregation-based POC devices to meet the signal and sensitivity needs for infectious disease diagnosis and other applications. KEYWORDS: nanoparticle aggregation, gold nanoparticle, point-of-care (POC) diagnosis, cross-linking aggregation, infectious diseases
I
considered to be a general framework for developing POC systems.3 Various bioassay platforms have been developed and
nfectious diseases such as malaria cause high mortality and there is a significant need for point-of-care (POC) diagnostic assays for infectious disease screening and early detection.1,2 The ASSURED criteria (affordable, sensitive, specific, userfriendly, rapid and robust, equipment-free, and deliverable) recommended by the World Health Organizations (WHO) is © 2017 American Chemical Society
Received: July 12, 2017 Accepted: October 10, 2017 Published: October 10, 2017 1627
DOI: 10.1021/acssensors.7b00482 ACS Sens. 2017, 2, 1627−1636
Article
ACS Sensors
GNP parameter based design guidelines for the colorimetric assay.
modified toward POC settings, such as paper-based diagnostics,4 isothermal nucleic acid amplification,5,6 and microfluidic ELISA.7 Despite such progress, one of the major scientific challenges for the biomedical research community is to develop, validate, and apply low cost, rapid, and ultrasensitive diagnostic tests for infectious diseases. Gold nanoparticle (GNP) colorimetric assay is a simple and fast detection method and is based on the aggregation of GNPs that results in a color change due to the plasmonic coupling.8 First developed using latex microparticles (i.e., latex agglutination assay9), the GNP-based colorimetric assay has advantages over microparticles including less nonspecific aggregation10 and better colorimetric visualization due to plasmon resonance and coupling.11 Upon aggregation, monodispersed GNPs change color from red to light purple. Due to its simplicity, GNP colorimetric assay has found broad applications including the detection of proteins,12,13 nucleic acids,6,14,15 viruses,16,17 heavy metals,18,19 aptamers,20,21 and bacteria.22,23 Nucleic acid detection using GNP colorimetric assay is frequently reported and has enabled detection of single base pair mismatch.24,25 Recent innovations have been attempted to improve the sensitivity and to simplify experimental conditions for nucleic acid detection using GNP aggregation assay. First, various readout methods have been tested to detect GNP aggregates including colorimetric scattering,11 linear scattering,26 surface enhanced Raman spectroscopy (SERS),27 and surface plasmon resonance microscopy (SPRM).28 These methods have been shown to improve the analytical sensitivity by 2 to 3 orders of magnitude. Second, the fact that hybridization reaction using GNPs can be performed in 30 min at room temperature using an optimized buffer with dextran sulfate has improved its potential for integration in a POC system.11,29 In terms of GNP parameters, it has been suggested that larger nanoparticles lead to better sensitivity.14 GNP parameters have been studied to understand various assays for detection of proteins,30,31 aptamers,20 amino acids,32 metal ions,33,34 and other molecular agents.35−37 However, it is unclear how GNP design parameters affect the colorimetric assay performance for nucleic acid detection. Furthermore, there is a lack of attempts to incorporate and test findings from the field of colloidal science to further improve the assay performance for nucleic acid detection. In this report, we systematically investigated the effects of GNP design parameters, including GNP size, concentration, and size combinations, with the goal of further improving the assay performance. Specifically, we evaluated the assay performance by colorimetric detection of malarial DNA and reported the limit of detection (LOD) and signal strength (magnitude of colorimetric peak shift). Our result confirms that larger GNP gives more sensitive detection and stronger signal. Furthermore, GNP concentration has a profound effect on the assay performance, where higher GNP probe concentrations lead to stronger overall signal and an optimal intermediate GNP concentration gives best LOD. Studies from the field of colloidal science suggest that nanoparticle aggregation is accelerated using mixture of small and large nanoparticles. Surprisingly, we did not observe any improvement in the assay performance when using GNP size combinations compared with using a homogeneous GNP size. Finally, we attempt to explain observed results and summarize key factors including nanoparticle aggregation rate, plasmonic coupling strength, and background signal that jointly determine the resulting colorimetric signal. Our study provides new knowledge on
■
MATERIALS AND METHODS
Materials. Analytical grade chemical reagents including gold(III) chloride trihydrate, sodium citrate, hydrochloric acid, nitric acid, hydroquinone, formamide, dextran sulfate, and sodium chloride were purchased from Sigma-Aldrich (St. Louis, MO, USA). The procured chemicals were used as received. HPLC purified oligonucleotides were purchased from Bio Basic Inc. (Markham, ON, Canada). 384 well plates were purchased from Thermo Scientific (Rochester, NY, USA). 2-Mercaptoethanol was procured from Fisher Scientific. Quant-iT Assay kit for GNP conjugated oligonucleotide quantification was purchased from Invitrogen. All experiments were performed with ultrapure water (Millipore, Billerica, MA, USA). GNP Synthesis and Characterization. All the glassware used in the GNP synthesis was cleaned using aqua regia (3:1 ratio of hydrochloric acid and nitric acid). 15 nm GNP was synthesized using Fren’s method with slight modifications.38,39 Briefly, gold chloride (III) solution was brought to a boil under continuous stirring, and then sodium citrate was added. After a prominent color change from purple to pink, the solution was cooled to room temperature. 30 nm and 50 nm particles were synthesized by hydroquinone reduction.40 Specifically, hydroquinone was added to a solution of gold chloride, sodium citrate, and the 15 nm GNPs (used as seed particles) at room temperature. This solution was stirred continually overnight to allow for GNP growth. The size and distribution of GNPs were characterized by dynamic light scattering (DLS) using Malvern Zetasizer Nano (Malvern Instruments Ltd., UK) and JEOL JEM 2100 transmission electron microscopy (TEM). GNP sample was spotted onto a thin carbon film coated Cu grids (300 mesh, Pacific Grid Tech) and air-dried for TEM characterization. TEM images were analyzed using ImageJ (National Institute of Health). GNP−Oligonucleotide Conjugation. Malaria (genus Plasmodium) was chosen as a clinically relevant target and the oligonucleotides sequences were designed complementary to the malarial sequence (Table S1).29 A poly A-tail was included in the oligonucleotides to increase flexibility of the sequence. The oligonucleotides were resuspended in ultrapure water to obtain required concentrations. The thiol-capped oligonucleotides A and B were conjugated separately to GNPs by a pH-assisted and surfactantfree method41 and the oligonucleotide: GNP ratios used was used (Table S2) to create stable GNP−oligonucleotide conjugates (Probe A and Probe B). Briefly, the pH of the mixture was adjusted to 3.0 ± 0.1 by the addition of citrate-HCl buffer (50 mM) in a 1:1 volumetric ratio. After 30 min of incubation, the GNP-oligo conjugates were centrifuged and washed with ultrapure water. The size and distribution of probes were characterized by DLS. The oligonucleotide loading density on the GNP surface was quantified using Oli-Green fluorescence assay by reducing the thiol bond following 18 h incubation with 1 M β-mercaptoethanol. The oligonucleotides A and B with known concentration were used as the calibration standard.42,43 The average loading density per GNP was measured and calculated for oligonucleotides A and B. Target Detection. The target sample was serially diluted to desired concentrations in terms of moles. The samples used in the assay ranged from 250 amol to 25 pmol for a target volume of 25 μL and the hybridization buffer was used as a blank. The freshly prepared probes were mixed with the hybridization buffer (20% formamide, 16% dextran sulfate, and 0.6 M sodium chloride) in a volume ratio of 3:3:4 (Probe A:Probe B:Buffer) to obtain the working solution.29 This working solution was then incubated with the target in a 2:1 volume ratio (working solution:target). UV−vis spectrum was measured in 384 well plates using a microplate reader (Biotek Synergy 2). All data was plotted using OriginLab. The LOD for each case was calculated using the IUPAC method,44 as detailed in the Supporting Information and Figure S1. Optical Properties Calculation of GNP and Aggregates Using DDA. Discrete Dipole Approximation (DDA) is a discrete 1628
DOI: 10.1021/acssensors.7b00482 ACS Sens. 2017, 2, 1627−1636
Article
ACS Sensors
Figure 1. Gold nanoparticle aggregation-based colorimetric assay. (a) Schematic of the DNA induced gold nanoparticle aggregation (at rate k) and its subsequent colorimetric detection (peak shift Δλ). The oligonucleotide conjugated gold nanoparticles (probe A and B) undergo hybridization in the presence of the target DNA to cause a plasmonic red shift and subsequent color change. The target sequence leads to head-to-tail clustering of GNPs. “c” refers the interparticle clearance distance. (b) UV−vis spectrum of the mixture of probes in the absence (control, black) and presence (target, red) of the target. Visual conformation of the target induced aggregation, i.e., color change (Inlet). (c) TEM images confirm the formation of nanoparticle aggregates (left: no target; right: with target). The TEM images were prepared using 30 nm GNP probes (peak absorbance = 0.75). solution method of the integral form of Maxwell’s equations and allows calculating nanostructure optical properties with complex geometries. Briefly, the target structure is discretized into a finite array of dipoles (N). After solving a large number (3N) of complex linear equations with unknown dipole moments, the extinction, absorption, and scattering cross sections can be predicted. In this study, the DDA package DDSCAT 7.3 developed by Draine and co-workers45 was implemented to solve the scattering and absorption properties of individual and aggregated GNPs. The dipole size effect on extinction efficiency factor in visible and near-infrared wavelength ranges was investigated for 15 nm GNP. Comparison with Mie theory and experimental data showed that a dipole size of 1 nm provides accurate response over the considered spectrum. The effect of nanoparticle size on the dielectric function was previously accounted for and adapted here. The dipole structures were visualized in an open-source software package ParaView. The extinction coefficient was calculated by using the following formula:
Q ext =
Table 1. Limit of Detection (LOD) Values for This Study
Figure 2b
2c,3b
2d
4b
Cext M
∑i = 1 πri2 4c
where Qext is the extinction efficiency factor, Cext is the extinction cross-section, and ri is the radius of individual GNP, and M is the number of GNPs in the aggregate (M ≥ 1). The denominator of the above equation refers to the summation of geometric cross-sectional areas for nanoparticles that form the aggregate, and is used here for convenient comparison with individual GNP.
■
d (nm)
absorbance
GNP concentration (pM)
15 30 50 15 30 50 15 30 50 30 30 30 30 30 50 50 50 50 50
0.055 0.5 2.5 0.75 0.75 0.75 0.3 0.4 0.55 0.15 0.4 0.5 0.75 1.85 0.15 0.55 0.75 1.85 2.5
150 150 150 2000 220 44 820 120 32 45 120 150 220 550 8.7 32 44 110 150
probe oligonucleotide concentration (nM) 9.3 56 213 126 85 64 51 45 47 17 45 56 85 208 13 47 64 158 213
LOD (fmol) b
54.2 22.5a 34 23.3 11.5 b
47.6 10.4 93.7 47.6 54.2 23.3 48.6 225a 10.4 11.5 23.6a 22.5a
a
This case has a very small or no dynamic range as shown in Figure 2c. This data does not have a linear range to allow calculation of the limit of detection (LOD).
b
RESULTS GNP−Oligonucleotide Conjugation and GNP Aggregation Assay. Conjugation of the oligonucleotide sequence on the GNP was confirmed by an increase in the GNP
hydrodynamic size (10−20 nm, Figure S2) and plasmonic resonance peak shift (5−20 nm, Figure S2). The conjugation 1629
DOI: 10.1021/acssensors.7b00482 ACS Sens. 2017, 2, 1627−1636
Article
ACS Sensors
Figure 2. Effect of gold nanoparticle size on the GNP-based aggregation assay performance. (a) Schematic of the change in nanoparticle size. The effect of GNP size was compared by keeping the GNP concentration (b, 150pM), peak absorbance (c, abs = 0.75) and the probe oligonucleotide concentration constant (d, 48 nM). Details calculations are available in Table S3.
Figure 3. Effect of gold nanoparticle size combinations on the aggregation assay performance. (a) Schematic of GNP size combinations in the study (all GNP have the same absorbance, abs = 0.75). (b) Peak shift for probes consisting of size combinations in comparison with homogeneous GNP sizes. (Solid lines: size combinations; dotted lines: homogeneous GNP probe size).
was performed using the molar ratios (oligonucleotide:GNP ratio) to obtain stable conjugates, as shown in Table S2. Furthermore, the number of oligonucleotides per GNP was measured using the Oli-Green fluorescence assay and was determined to be approximately 62 ± 14, 379 ± 112, and 1467 ± 480 for 15, 30, and 50 nm GNPs, respectively. GNP aggregation due to DNA hybridization was confirmed by the plasmonic peak shift using the UV−vis spectrum, a visible color change, and TEM (Figure 1, more TEM images in Figures S3 and S4). No peak shift was observed in the absence of target sequence or with control target sequence (Table S1). TEM
images were obtained before and after the introduction of target DNA (Figure 1c). TEM image analysis shows an interparticle distance c ≈ 5 nm (Figure S5). This distance was further used as a reference for computational simulation by DDA. Effect of Nanoparticle Size. Three GNP sizes were studied including 15, 30, and 50 nm with peak absorbance at 519, 525, and 535 nm, respectively (Figure S2). We systematically investigated the effect of nanoparticle size by separately keeping the following parameters constant: (1) GNP concentration, (2) UV−vis peak absorbance of GNP1630
DOI: 10.1021/acssensors.7b00482 ACS Sens. 2017, 2, 1627−1636
Article
ACS Sensors
Figure 4. Effect of gold nanoparticle concentration on the aggregation assay performance. (a) The effect of GNP concentration was investigated for 30 and 50 nm size gold nanoparticles. (b) Peak shift for 30 nm GNP probes at different GNP concentrations (45, 120, 150, 220, and 550 pM). (c) Peak shift for 50 nm probes at different GNP concentrations (8.7, 32, 44, 110, 150 pM).
Figure 5. Numerical simulation of the plasmonic coupling between two gold nanoparticles. Two GNPs are separated by distance “c” for different size combinations including (a) 15/15 nm; (b) 30/30 nm; (c) 50/50 nm; (d) 15/30 nm; (e) 15/50 nm; and (f) 30/50 nm. (a−c) also include spectrum for single GNP as represented by the dotted blue line.
oligonucleotide conjugate, and (3) the probe oligonucleotide concentration. The corresponding molar concentration and absorbance values are detailed in Table 1. To quantitatively compare the colorimetric assay performance, we evaluated the plasmonic peak shift at 30 min and calculated the peak shift by
evaluating the peak absorbance of the target and comparing with that of the control. The difference between these two peak wavelengths was considered to be the peak shift (Figure 1b). Changing the nanoparticle size has a significant impact on the colorimetric assay performance including the limit of 1631
DOI: 10.1021/acssensors.7b00482 ACS Sens. 2017, 2, 1627−1636
Article
ACS Sensors
Figure 6. Numerical simulation of the plasmonic coupling between eight gold nanoparticles. Eight GNPs positioned in a simple cubic lattice. (a) Schematic of GNP positions for c = 1 nm (top) and c = 5 nm (bottom); Spectrum for 15 nm (b), 30 nm (c), and 50 nm (d) GNP aggregates at interparticle distance of 1 nm (orange) and 5 nm (blue), in comparison with single GNP (solid black line).
detection and signal strength (i.e., magnitude of the peak shift). First, with constant nanoparticle concentration (150 pM, Figure 2b), increasing nanoparticle size leads to higher signal strength and improved (lower) LOD. Specifically, LOD decreases from 54.2 fmol for 30 nm GNP to 22.5 fmol for 50 nm GNP. The peak shift for 15 nm GNP was very small and did not yield a reliable estimation of LOD. It should be noted that the GNP absorbance values are very different when the nanoparticle concentration are kept constant, ranging from 2.5 for 50 nm, 0.5 for 30 nm, to 0.055 for 15 nm. Next, when keeping the absorbance of the nanoparticles consistent, 30 and 50 nm GNPs have similar magnitude in the peak shift (Figure 2c). Larger GNP size leads to improved LOD values as detailed in Table 1. Finally, under the same probe oligonucleotide concentration (Figure 2d), 30 and 50 nm GNP have higher peak shift and an improved LOD for 50 nm GNP. In all cases, larger GNP size leads to stronger signal strength and improved LOD (Figure 7, case 1). All LOD values for Figure 2 are also listed in Table 1. Effect of GNP Size Combinations. We then investigated the effect of combining two different GNP sizes for the colorimetric assay. The absorbance of GNP probe was kept constant (abs = 0.75). The size combinations include 15/30 nm, 15/50 nm, and 30/50 nm (Figure 3a). Overall, GNP size
combinations did not improve the sensitivity or signal strength of the assay when compared to homogeneous GNP sizes (Figure 3b). 15/30 nm combinations gave slightly better signal strength when compared to homogeneous 15 nm GNPs. 30/50 nm combination gave slightly higher peak shift than 15/50 nm and 15/30 nm combinations, but did not outperform the homogeneous GNP sizes (30 and 50 nm). LOD values for GNP size combinations are significantly higher than homogeneous GNP size groups (Figure 7, case 2) and are listed in Table S3. Effect of Nanoparticle Concentration. We next investigated the effect of GNP concentration on the colorimetric assay performance (Figure 4). We chose 30 and 50 nm GNP since 15 nm did not yield a significant peak shift. For both 30 and 50 nm GNPs, higher GNP probe concentration leads to larger peak shift. For LOD, it appears that there is an optimal GNP concentration that lead to the lowest LOD, specifically 220 pM for 30 nm GNP (Table 1, Figure 7, case 3) and 32 pM for 50 nm GNP (Figure S6). Simulation of the Plasmonic Coupling-Induced Peak Shift. We used discrete dipole approximation (DDA) to calculate the extinction spectrum of GNP aggregates. To illustrate the underlying physics and to keep a reasonable simulation domain, we studied the plasmonic coupling of two 1632
DOI: 10.1021/acssensors.7b00482 ACS Sens. 2017, 2, 1627−1636
Article
ACS Sensors
Figure 7. Physical mechanisms that determine the colorimetric assay performance. (a) Schematic of key factors that determine the colorimetric assay peak shift, including the aggregation rate (k), plasmonic coupling strength (g), and signal versus background absorption. (b) Effect of nanoparticle size: summary of the contributing factors including (i) aggregation rate and (ii) plasmonic coupling strength as a function of nanoparticle size, and effect of nanoparticle size (iii, case 1) and size combinations (iv, case 2) on the peak shift and LOD. (c) Effect of nanoparticle concentration: summary of contributing factors including (i) aggregation rate and (ii) signal versus background absorption as a function of nanoparticle concentration, and effect of nanoparticle concentration on the peak shift (iii, case 3). Case 1: peak shift was plotted for 790 fmol from Figure 2b−d; Case 2: peak shift was plotted for 790 fmol from Figure 3b; Case 3: peak shift was plotted for 790 fmol from Figure 4a, 30 nm GNP.
■
DISCUSSION Our study focuses on effects of GNP size, concentration, and size combinations on the colorimetric assay performance. Analysis of the results leads us to elucidate three key factors that determine the colorimetric assay performance, including the aggregation rate, the strength of plasmonic coupling between aggregated GNPs, and the background signal (Figure 7a). To generate a colorimetric peak shift, nanoparticles need to diffuse and interact with other particles for the hybridization and thus aggregation. The aggregation rate may increase with higher nanoparticle concentration due to higher possibility of nanoparticle interactions, decrease with nanoparticle size due to slower diffusion, and increase by mixing different nanoparticle sizes.46 After aggregation, nanoparticle plasmons couple between aggregated nanoparticles by Coulomb interactions of opposite charges,47 and leads to a shifted UV−vis absorption, or colorimetric peak shift. Increasing nanoparticle size leads to much stronger plasmonic coupling (Figures 5 and 6). Finally, the colorimetric peak shift among the aggregated nanoparticle population has to overcome the background absorption signal from nonaggregated nanoparticles to generate a measurable
GNPs (Figure 5) and eight GNPs (Figure 6). GNPs were positioned at different clearances (c = 0, 1, 3, and 5 nm) and the extinction spectrum was calculated by averaging different incident light directions. Plasmonic coupling of two GNPs generates a second peak at longer wavelength (650 to 750 nm) in addition to the primary peak (500 to 550 nm). Comparing the plasmonic coupling of two identically sized GNPs, the magnitude of the second peak for 50 nm is much larger than 15 and 30 nm GNPs, and persists when increasing the clearance distance (Figure 5a-c). This observation was also validated while simulating eight GNPs which were positioned on a simple cubic lattice cell (Figure 6). Furthermore, the plasmonic coupling for size combinations is weaker than the two homogeneous GNPs and is limited by the smaller GNP size in the size mixture. For instance, the 15/30 nm combination gives a smaller second peak when compared with both homogeneous 15 and 30 nm cases (Figure 5d versus a and b). The 15/50 nm combination gives a very small second peak even when the two GNPs are in contact (c = 0 nm, Figure 5e). 1633
DOI: 10.1021/acssensors.7b00482 ACS Sens. 2017, 2, 1627−1636
Article
ACS Sensors
strength. Higher GNP concentration leads to stronger signal at high target concentrations as a result of higher aggregation rate; however, the fact that the optimal LOD was found at intermediate GNP concentrations suggests a balance of two competing mechanisms including aggregation rate and signal/ background ratio. This work contributes new knowledge for better design of colorimetric assays for infectious disease POC diagnosis and other applications.
color shift. First, we systematically examined the effect of GNP size and found that increasing nanoparticle size leads to higher overall signal strength, i.e., the magnitude of the peak shift, and improved LOD (Figure 7b and case 1). From a physical point of view, larger particles diffuse slower and may give a slower aggregation rate (Figure 7b, (i)).46 However, the plasmonic coupling strength is much higher for larger nanoparticles (Figures 5 and 6, Figure 7b (ii)). The fact that larger GNP gives higher colorimetric peak shift suggests a dominant contribution of the plasmonic coupling strength over aggregation rate. Second, previous studies from the field of colloidal science suggest that nanoparticle aggregation is accelerated using a mixture of small and large nanoparticles, i.e., a binary system.46 The smaller particles move toward a single large particle in the binary system to form stable aggregates. We tested using two GNP sizes for each probe (A and B, Figure 4) and evaluated the assay performance. In comparison to homogeneous GNP sizes, the size combinations fail to improve the assay performance. This led us to hypothesize that while the size combination may accelerate the aggregation rate, the plasmonic coupling between two different GNP sizes is weaker based on our numerical analysis (Figure 5; Figure 7b and case 2). As a result, the size combination does not lead to improvement in the overall assay performance. This again suggests the important role of plasmonic coupling strength over the aggregation rate for nanoparticle size mixtures. Third, as compared to GNP size, less is known on how the GNP probe concentration affects the target detection. Our findings suggest that GNP probe concentration has a profound impact on the colorimetric assay performance, with higher GNP probe concentrations leading to stronger overall signal. Furthermore, an optimal intermediate GNP concentration gives best LOD (Figure 7, case 3 for 30 nm GNP; Figure S6 for 50 nm GNP). Varying GNP concentration involves two competing physical mechanisms: (1) higher probability of GNP interaction and thus accelerated GNP aggregation rate with higher GNP concentrations (Figure 7c, (i)); and (2) higher background signal from excess nonaggregated GNP at high GNP concentrations (thus lower signal/background ratio, Figure 7c, (ii)). The fact that higher GNP concentration gives higher peak shift suggests a dominating effect of GNP aggregation rate at high target concentrations. On the other hand, the optimal LOD at intermediate GNP concentration suggests the balance of the two competing effects at low target concentrations. Finally, we observed the signal drop at very high target concentrations (Figures 2 to 4) due to a phenomenon similar to the hook effect or Prozone effect, previously observed in immunoassays.48,49 This is due to the saturation of the GNP probe caused by the excess of target DNA available in the solution. This saturation effect makes them unavailable for cross-linking (Figure 1) and thus reduces aggregation and signal. The onset of this effect depends on the total number of probe oligonucleotides available. For instance, at the same GNP concentration (Figure 2b), this effect takes place at a much lower target concentration for 30 nm GNP than 50 nm GNP, due to less probe oligonucleotide number on the 30 nm GNP. In conclusion, our study elucidates the three contributing factors that determine the colorimetric assay performance including the nanoparticle aggregation rate, the strength of plasmonic coupling between nanoparticles, and background signal. Larger GNP size leads to stronger signal and improved LOD, suggesting a dominating role of plasmonic coupling
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acssensors.7b00482. Calculations for GNP molar concentration and Limit of Detection (LOD); Details on the oligonucleotides and target sequences, conjugation criteria, and LOD values; UV−vis characterization of synthesized GNP, their conjugation, aggregation, and reaction kinetics for 15, 30, and 50 nm GNPs; TEM images indicating presence and absence of 30 and 50 nm GNP aggregation; histogram plot for interparticle clearance distance; summarized effect of GNP concentration on the peak shift and LOD for 50 nm GNP (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
Zhenpeng Qin: 0000-0003-3406-3045 Author Contributions #
Co-first authors.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors thank Kyryl Zagorovsky from Dr. Warren Chan’s group at University of Toronto for helpful discussions and suggesting the pH-assisted conjugation approach. This study is supported by a Texas Medical Research Collaborative (TxMRC) grant and startup fund from The University of Texas at Dallas (UTD). The authors acknowledge HPC resources from the Texas Advanced Computing Center (TACC) at The University of Texas at Austin, the ROMEO HPC Center hosted by the University of Reims ChampagneArdenne, and the Extreme Science and Engineering Discovery Environment (XSEDE) which is supported by National Science Foundation grant number ACI-1053575. J.R. thanks the University of Texas at Dallas for supporting his visit at the Z.Q.’s group at UTD.
■
REFERENCES
(1) Sun, J.; Xianyu, Y.; Jiang, X. Point-of-care biochemical assays using gold nanoparticle-implemented microfluidics. Chem. Soc. Rev. 2014, 43, 6239−6253. (2) Hauck, T. S.; Giri, S.; Gao, Y.; Chan, W. C. W. Nanotechnology diagnostics for infectious diseases prevalent in developing countries. Adv. Drug Delivery Rev. 2010, 62, 438−448. (3) Abou Tayoun, A. N.; Burchard, P. R.; Malik, I.; Scherer, A.; Tsongalis, G. J. Democratizing Molecular Diagnostics for the Developing World. Am. J. Clin. Pathol. 2014, 141, 17−24.
1634
DOI: 10.1021/acssensors.7b00482 ACS Sens. 2017, 2, 1627−1636
Article
ACS Sensors (4) Wang, Y.; Li, D.; Ren, W.; Liu, Z.; Dong, S.; Wang, E. Ultrasensitive colorimetric detection of protein by aptamer-Au nanoparticles conjugates based on a dot-blot assay. Chem. Commun. (Cambridge, U. K.) 2008, 22, 2520−2522. (5) Chen, F.; Zhao, Y.; Fan, C.; Zhao, Y. Mismatch Extension of DNA Polymerases and High-Accuracy Single Nucleotide Polymorphism Diagnostics by Gold Nanoparticle-Improved Isothermal Amplification. Anal. Chem. 2015, 87, 8718−8723. (6) Ravan, H. Isothermal RNA detection through the formation of DNA concatemers containing HRP-mimicking DNAzymes on the surface of gold nanoparticles. Biosens. Bioelectron. 2016, 80, 67−73. (7) Ng, A. H. C.; Uddayasankar, U.; Wheeler, A. R. Immunoassays in microfluidic systems. Anal. Bioanal. Chem. 2010, 397, 991−1007. (8) Storhoff, J.; Elghanian, R.; Mucic, R.; Mirkin, C.; Letsinger, R. One-pot colorimetric differentiation of polynucleotides with single base imperfections using gold nanoparticle probes. J. Am. Chem. Soc. 1998, 120, 1959−1964. (9) Smits, H. L.; van der Hoorn, M. A.; Goris, M. G.; Gussenhoven, G. C.; Yersin, C.; Sasaki, D. M.; Terpstra, W. J.; Hartskeerl, R. A. Simple latex agglutination assay for rapid serodiagnosis of human leptospirosis. J. Clin. Microbiol. 2000, 38, 1272−1275. (10) Li, H.; Rothberg, L. Colorimetric detection of DNA sequences based on electrostatic interactions with unmodified gold nanoparticles. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 14036−14039. (11) Storhoff, J.; Lucas, A.; Garimella, V.; Bao, Y.; Muller, U. Homogeneous detection of unamplified genomic DNA sequences based on colorimetric scatter of gold nanoparticle probes. Nat. Biotechnol. 2004, 22, 883−887. (12) Nam, J.; Thaxton, C.; Mirkin, C. Nanoparticle-based bio-bar codes for the ultrasensitive detection of proteins. Science 2003, 301, 1884−1886. (13) Li, J.; Fu, H.; Wu, L.; Zheng, A.; Chen, G.; Yang, H. General Colorimetric Detection of Proteins and Small Molecules Based on Cyclic Enzymatic Signal Amplification and Hairpin Aptamer Probe. Anal. Chem. 2012, 84, 5309−5315. (14) Reynolds, R.; Mirkin, C.; Letsinger, R. Homogeneous, nanoparticle-based quantitative colorimetric detection of oligonucleotides. J. Am. Chem. Soc. 2000, 122, 3795−3796. (15) Rajendran, P.; Kaufmann, S.; Voeroes, J.; Zenobi-Wong, M.; Demko, L. Femtomolar oligonucleotide detection by a one-step gold nanoparticle-based assay. Colloids Surf., B 2015, 135, 193−200. (16) Saleh, M.; El-Matbouli, M. Rapid detection of Cyprinid herpesvirus-3 (CyHV-3) using a gold nanoparticle-based hybridization assay. J. Virol. Methods 2015, 217, 50−54. (17) Dharanivasan, G.; Riyaz, S. U. M.; Jesse, D. M. I.; Muthuramalingam, T. R.; Rajendran, G.; Kathiravan, K. DNA templated self-assembly of gold nanoparticle clusters in the colorimetric detection of plant viral DNA using a gold nanoparticle conjugated bifunctional oligonucleotide probe. RSC Adv. 2016, 6, 11773−11785. (18) Lin, Y. W.; Huang, C. C.; Chang, H. T. Gold nanoparticle probes for the detection of mercury, lead and copper ions. Analyst 2011, 136, 863−871. (19) Kanayama, N.; Takarada, T.; Maeda, M. Rapid naked-eye detection of mercury ions based on non-crosslinking aggregation of double-stranded DNA-carrying gold nanoparticles. Chem. Commun. (Cambridge, U. K.) 2011, 47, 2077−2079. (20) Gopinath, S. C. B.; Lakshmipriya, T.; Awazu, K. Colorimetric detection of controlled assembly and disassembly of aptamers on unmodified gold nanoparticles. Biosens. Bioelectron. 2014, 51, 115−123. (21) Du, G.; Zhang, D.; Xia, B.; Xu, L.; Wu, S.; Zhan, S.; Ni, X.; Zhou, X.; Wang, L. A label-free colorimetric progesterone aptasensor based on the aggregation of gold nanoparticles. Microchim. Acta 2016, 183, 2251−2258. (22) Khan, S. A.; DeGrasse, J. A.; Yakes, B. J.; Croley, T. R. Rapid and sensitive detection of cholera toxin using gold nanoparticle-based simple colorimetric and dynamic light scattering assay. Anal. Chim. Acta 2015, 892, 167−174.
(23) Verma, M. S.; Rogowski, J. L.; Jones, L.; Gu, F. X. Colorimetric biosensing of pathogens using gold nanoparticles. Biotechnol. Adv. 2015, 33, 666−680. (24) Fang, W.; Chen, W.; Yang, J. Colorimetric determination of DNA concentration and mismatches using hybridization-mediated growth of gold nanoparticle probes. Sens. Actuators, B 2014, 192, 77− 82. (25) Zhan, Z.; Ma, X.; Cao, C.; Sim, S. J. Gold-based optical biosensor for single-mismatched DNA detection using salt-induced hybridization. Biosens. Bioelectron. 2012, 32, 127−132. (26) Du, B.; Li, Z.; Liu, C. One-step homogeneous detection of DNA hybridization with gold nanoparticle probes by using a linear lightscattering technique. Angew. Chem., Int. Ed. 2006, 45, 8022−8025. (27) Frost, M. S.; Dempsey, M. J.; Whitehead, D. E. Highly sensitive SERS detection of Pb2+ ions in aqueous media using citrate functionalised gold nanoparticles. Sens. Actuators, B 2015, 221, 1003−1008. (28) Yuan, L.; Wang, X.; Fang, Y.; Liu, C.; Jiang, D.; Wo, X.; Wang, W.; Chen, H. Digitizing Gold Nanoparticle-Based Colorimetric Assay by Imaging and Counting Single Nanoparticles. Anal. Chem. 2016, 88, 2321−2326. (29) Cordray, M. S.; Amdahl, M.; Richards-Kortum, R. R. Gold nanoparticle aggregation for quantification of oligonucleotides: optimization and increased dynamic range. Anal. Biochem. 2012, 431, 99−105. (30) Thanh, N.; Rosenzweig, Z. Development of an aggregation based immunoassay for anti protein a using gold nanoparticles. Abstracts of Papers of the American Chemical Society 2002, 223, U74− U74. (31) Liu, X.; Wang, Y.; Chen, P.; McCadden, A.; Palaniappan, A.; Zhang, J.; Liedberg, B. Peptide Functionalized Gold Nanoparticles with Optimized Particle Size and Concentration for Colorimetric Assay Development: Detection of Cardiac Troponin I. Acs Sensors 2016, 1, 1416−1422. (32) Li, L.; Li, B. Sensitive and selective detection of cysteine using gold nanoparticles as colorimetric probes. Analyst 2009, 134, 1361− 1365. (33) Liu, J.; Lu, Y. Accelerated color change of gold nanoparticles assembled by DNAzymes for simple and fast colorimetric Pb2+ detection. J. Am. Chem. Soc. 2004, 126, 12298−12305. (34) Krpetic, Z.; Guerrini, L.; Larmour, I. A.; Reglinski, J.; Faulds, K.; Graham, D. Importance of Nanoparticle Size in Colorimetric and SERS-Based Multimodal Trace Detection of Ni(II) Ions with Functional Gold Nanoparticles. Small 2012, 8, 707−714. (35) Medley, C. D.; Smith, J. E.; Tang, Z.; Wu, Y.; Bamrungsap, S.; Tan, W. Gold nanoparticle-based colorimetric assay for the direct detection of cancerous cells. Anal. Chem. 2008, 80, 1067−1072. (36) Chen, Z.; Wang, Z.; Chen, J.; Wang, S.; Huang, X. Sensitive and selective detection of glutathione based on resonance light scattering using sensitive gold nanoparticles as colorimetric probes. Analyst 2012, 137, 3132−3137. (37) Poonthiyil, V.; Golovko, V. B.; Fairbanks, A. J. Size-optimized galactose-capped gold nanoparticles for the colorimetric detection of heat-labile enterotoxin at nanomolar concentrations. Org. Biomol. Chem. 2015, 13, 5215−5223. (38) Kimling, J.; Maier, M.; Okenve, B.; Kotaidis, V.; Ballot, H.; Plech, A. Turkevich method for gold nanoparticle synthesis revisited. J. Phys. Chem. B 2006, 110, 15700−15707. (39) Frens, G. Controlled Nucleation for Regulation of Particle-Size in Monodisperse Gold Suspensions. Nature, Phys. Sci. 1973, 241, 20− 22. (40) Perrault, S. D.; Chan, W. C. Synthesis and surface modification of highly monodispersed, spherical gold nanoparticles of 50−200 nm. J. Am. Chem. Soc. 2009, 131, 17042−17043. (41) Zhang, X.; Servos, M. R.; Liu, J. Instantaneous and quantitative functionalization of gold nanoparticles with thiolated DNA using a pH-assisted and surfactant-free route. J. Am. Chem. Soc. 2012, 134, 7266−7269. 1635
DOI: 10.1021/acssensors.7b00482 ACS Sens. 2017, 2, 1627−1636
Article
ACS Sensors (42) Demers, L.; Mirkin, C.; Mucic, R.; Reynolds, R.; Letsinger, R.; Elghanian, R.; Viswanadham, G. A fluorescence-based method for determining the surface coverage and hybridization efficiency of thiolcapped oligonucleotides bound to gold thin films and nanoparticles. Anal. Chem. 2000, 72, 5535−5541. (43) Hill, H. D.; Millstone, J. E.; Banholzer, M. J.; Mirkin, C. A. The Role Radius of Curvature Plays in Thiolated Oligonucleotide Loading on Gold Nanoparticles. ACS Nano 2009, 3, 418−424. (44) Long, G. L.; Winefordner, J. D. Limit of detection. A closer look at the IUPAC definition. Anal. Chem. 1983, 55, 712A−724A. (45) Draine, B.; Flatau, P. Discrete-Dipole Approximation for Scattering Calculations. J. Opt. Soc. Am. A 1994, 11, 1491−1499. (46) Costanzo, P.; Patten, T.; Seery, T. Nanoparticle agglutination: Acceleration of aggregation rates and broadening of the analyte concentration range using mixtures of various-sized nanoparticles. Langmuir 2006, 22, 2788−2794. (47) Nordlander, P.; Oubre, C.; Prodan, E.; Li, K.; Stockman, M. Plasmon hybridizaton in nanoparticle dimers. Nano Lett. 2004, 4, 899−903. (48) Gillet, P.; Mori, M.; Van Esbroeck, M.; Van den Ende, J.; Jacobs, J. Assessment of the Prozone effect in malaria rapid diagnostic tests. Malar. J. 2009, 8, 271. (49) Schiettecatte, J.; Anckaert, E.; Smitz, J. In Interferences in Immunoassays; Chiu, N., Ed.; Advances in Immunoassay Technology; InTech, 2012; Vol. 1, p 45.
1636
DOI: 10.1021/acssensors.7b00482 ACS Sens. 2017, 2, 1627−1636