Sensitive Detection of Platelet-Derived Growth Factor through Surface

Article pubs.acs.org/ac

Sensitive Detection of Platelet-Derived Growth Factor through Surface-Enhanced Raman Scattering Chia-Wei Wang and Huan-Tsung Chang* Department of Chemistry, National Taiwan University, 1, Section 4, Roosevelt Road, Taipei 10617, Taiwan S Supporting Information *

ABSTRACT: A surface-enhanced Raman scattering (SERS) assay using two different nanomaterials has been demonstrated for highly sensitive and selective detection of platelet-derived growth factor (PDGF). Gold nanoparticles (Au NPs; 13 nm) are conjugated with aptamer (Apt) and 4-mercaptobenzoic acid (MBA) as the recognition element and reporter, respectively, while Au pearl necklace nanomaterials (Au PNNs) are used for generating reproducible and enhanced SERS signal of 4-MBA. The Apt/MBA−Au NPs bind PDGF through a specific interaction between Apt and PDGF in a fashion of 2:1, leading to concentration of the analyte and removal of the sample matrix. Through electrostatic interaction, the PDGF−Apt/ MBA−Au NPs complexes form aggregates with Au PNNs, leading to an enhanced Raman signal of 4-MBA. Au PNNs allow enhancement factors up to 1.3 × 107 and relative standard deviations of Raman signals for 4-MBA down to 15% (five measurements). The assay allows detection of PDGF BB down to 0.5 pM, with linearity of the Raman signal of 4-MBA against the concentration of PDGF over 1−50 pM. Having advantages of sensitivity and reproducibility, this assay has been further applied for the determination of the concentration of PDGF in urine samples, showing its great potential for ultrasensitive analysis of target proteins in biological samples.

N

applied for sensing of other important analytes such as plateletderived growth factors (PDGF). PDGFs are proteins found in human platelets that regulate the tumor growth and division through their binding to the tyrosine receptors on the cell membrane.18 They are dimeric molecules, each consisting of disulfide-bonded, structurally similar A- and Bpolypeptide chains.19 Overexpression of PDGFs has been implicated in nonmalignant conditions characterized by an increased cell proliferation, such as atherosclerosis and fibrotic conditions.20 As a potential marker of cancer diagnosis, several PDGF sensing methods have been reported, including colorimetric, luminescence, fluorescence, and electrochemical methods.21−24 Aptamers (Apt) conjugated Au NPs have been demonstrated for the detection of PDGF down to 80 and 300 pM through fluorescence and absorption detection modes, respectively. The Apt with a sequence of 5′-CAG GCT ACG GCA CGT AGA GCA TCA CCA TGA TCC TG-3′ binds with PDGF BB in a fashion of 2:1 with high affinity.22 Although they are sensitive, it is still difficult to detect PDGF in real samples, mainly because of their trace amounts in the presence of abundant proteins and other small interfering species. Thus, techniques providing high sensitivity and specificity are still highly demanded.

anomaterials such as gold (Au) and silver (Ag) nanoparticles (NPs) have been employed to develop sensitive optical detection systems for various analytes, including proteins, DNA, and small molecules.1−3 Relative to Au NPs, Ag NPs providing enhancement factors (EF) up to 1014 at a hot spot are more commonly used in surface-enhanced Raman scattering (SERS).4−7 Having advantages of sensitivity and providing structural information, SERS has become one of the most popular sensing systems for various analytes.8−11 Although SERS using Ag NPs can provide great sensitivity, instability of Ag NPs due to oxidation is problematic. In addition, relative standard deviations (RSDs) of SERS signals are usually >20%, due to inhomogeneous distribution of Ag NPs aggregates on the substrate surface.12 Thus, there is still a need to develop sensitive and reproducible SERS systems. Relative to Ag NPs, Au NPs are more stable and easily prepared in high quality, but their enhancement factors are about 100− 1000 times lower.13,14 SERS using Au NPs has been employed for sensing of trinitrotoluene, with EF values up to 109.15 However, reproducibility of SERS using Au NPs as that using Ag NPs is not great. Pearl necklace nanomaterials (Au PNNs) were used in SERS to provide EF values of rhodamine 6G up to 5.6 × 109 and RSD < 10%.16 The SERS assay using Au PNNs allowed quantitative detection of human serum albumin (HSA) in urine, with a limit of detection (LOD) of 70 pM at a signal-tonoise ratio of 3.17 Although the assay allowed detection of HSA by taking advantages of its high specificity for AB 580, it cannot be © 2014 American Chemical Society

Received: April 18, 2014 Accepted: July 3, 2014 Published: July 3, 2014 7606

dx.doi.org/10.1021/ac5014207 | Anal. Chem. 2014, 86, 7606−7611

Analytical Chemistry

Article

phosphate buffer (10 mM, pH 7.0) containing as-prepared Au PNNs (1×). After 30 min, the solutions were subjected to centrifugation (6000g, 10 min) and then the pellets were resuspended in ultrapure water (10 μL). Finally, drops of the solutions were added separately onto silicon wafers and dried at room temperature. Urine samples from a healthy volunteer were filtered through a 0.22 μm membrane and stored at 4 °C. When required for use, the samples were thawed at ambient temperature. Standard PDGF solutions (10 μL, final concentration 5−30 pM) were then spiked into the sample solution (100 μL). The spiked urine samples were further diluted with phosphate solution (final concentration 10 mM, pH 7.0) to 0.5 mL. The diluted urine samples were analyzed in the same manner as that for the standard PDGF solutions. Triplicate measurements were conducted for each sample.

In this study, a new SERS approach in conjunction with Apt/4mercaptobenzoic acid (MBA) conjugated Au NPs (Apt/MBA− Au NPs) and Au PNNs was developed. Apt/MBA−Au NPs were used as a selector and reporter, while Au PNNs were used as a signal amplifier. Important factors such as the concentration of 4MBA on Apt/MBA−Au NPs and the concentration of Apt/ MBA−Au NPs were evaluated. The sensitive and selective SERS assay was validated by the analysis of PDGFs in urine, showing its great potential for cancer diagnosis.

■

EXPERIMENTAL SECTION Materials. Thiol-modified PDGF-binding aptamer (5′-HST30 CAG GCT ACG GCA CGT AGA GCA TCA CCA TGA TCC TG-3′) was purchased from Integrated DNA Technology, Inc. (Coralville, IA, U.S.A.). Recombinant human PDGF AA, PDGF AB, and PDGF BB were purchased from R&D Systems (Minneapolis, MN, U.S.A.). L-Cysteine, hexadecyltrimethylammonium bromide (CTAB), and sodium dodecyl sulfate (SDS) were obtained from Acros (Geel, Belgium). Sodium tetrachloroaurate dihydrate and trisodium citrate dihydrate were obtained from Aldrich (Milwaukee, WI, U.S.A.). Adenosine triphosphate (ATP), cytochrome c, L-glutathione (GSH), hemoglobin, myoglobin, and tripsinogen were purchased from Sigma (St. Louis, MO, U.S.A.). Lysozyme was purchased from MP Biomedicals LLC (Santa Ana, CA, U.S.A.). Hydrazine monohydrate and tellurium dioxide were purchased from SHOWA (Tokyo, Japan). Sodium phosphate dibasic anhydrous and sodium phosphate monobasic monohydrate were purchased from JT Baker (Phillipsburg, NJ, U.S.A.). Ultrapure water was obtained from a Milli-Q ultrapure (18.2 MΩ·cm) system. Synthesis of Au PNNs and Apt−Au NPs. Tellurium dioxide (0.016 g) and N2H4 (10 mL, 80%) as a reducing agent were used to prepare Te nanowires that were further used to prepare Au PNNs with NaAuCl4 (1 mL, 1 mM) through the galvanic reaction between Au3+ and Te.25 For simplicity, the concentration of as-prepared Au PNNs is denoted as 1×. Thiolmodified Apt (200 μM, 10 μL), 4-MBA (200 μM, 10 μL), and 13 nm Au NPs (15 nM, 980 μL) that were prepared from NaAuCl4 (1 mM, 250 mL) and 38.8 mM trisodium citrate (25 mL) were used to prepare Apt/MBA−Au NPs.26 Through Au−S bonding, Apt and 4-MBA molecules were bound to Au NPs. After 8 h, the mixture was centrifuged at a relative centrifugation force (RCF) of 30 000g for 25 min to remove the excess Apt and 4-MBA. After three cycles of centrifuge/wash (10 mL of ultrapure water × 3), the pellet was resuspended in Tris−HCl (5.0 mM, pH 7.4, 980 μL) and stored in a refrigerator (4 °C). Characterization. JEOL-1200EX II and FEI Tecnai-G2-F20 transmission electron microscopes were used to measure the sizes and shapes of the Au NPs and Au PNN. The ξ-potentials were measured on a Zetasizer 3000HS analyzer (Malvern Instruments, Malvern, U.K.). Raman spectra were obtained using a dispersive Raman spectrometer (ProTrustTech, Tainan, Taiwan) equipped with a 50× objective, a microscope, and a charge-coupled detector. The excitation wavelength was set at 532 nm and the signal collection time was 30 s. SERS Analysis of PDGFs. Aliquots of Apt/MBA−Au NPs solution (3 nM with respect to Au NPs, 50 μL) were separately added to phosphate solutions (950 μL, pH 7.0 10 mM) containing PDGF (from 1.05 to 52.6 pM). After being equilibrated at ambient temperature (25 °C) for 30 min, the mixtures were separately subjected to centrifugation (30 000g, 20 min). After two cycles of centrifuge/wash (ultrapure water, 1 mL × 2), the pellets were separately resuspended in 1 mL of

■

RESULTS AND DISCUSSION Sensing Strategy. Scheme 1 reveals the SERS sensing strategy toward PDGFs. Apt/MBA−Au NPs act as a recognition

Scheme 1. Schematic Presentation of the SERS Assay for the Sensing of PDGFs

element and a reporter. Through specific binding of PDGFs with Apt, PDGFs induce the formation of PDGF−Apt/MBA−Au NPs aggregates. The redispersed PDGF−Apt/MBA−Au NPs interact with Au PNNs to form aggregates through the electrostatic interaction between PDGF and Au PNNs. The Au PNNs act as a signal amplifier to enhance the Raman signal of 4MBA on the surface of Apt/MBA−Au NPs. Upon increasing PDGF concentration, greater amounts of Apt/MBA−Au NPs are bound to Au PNNs, leading to increased SERS signals of 4-MBA. The transmission electron microscopy (TEM) image (Figure 1C) displays the formation of Apt/MBA−Au NPs and Au PNNs aggregates in the presence of PDGF BB (50 pM). At such a low concentration, PDGF BB did not induce the formation of precipitate of Apt/MBA−Au NPs with Au PNNs. As a result, only a slight red shift of the surface plasmon resonance (SPR) absorption band at 530 nm was observed as shown in Figure 1D. We note that PDGF BB at 50 pM did not induce any shift of the SPR band of Apt/MBA−Au NPs in the absence of Au PNNs. At pH 7.0, PDGFs (pI 9.8) are positive, while Au PNNs and Apt/ MBA−Au NPs are both negative because they separately contained SDS and Apt/4-MBA (pKa 4.7) anions on their surfaces.27 In the absence of PDGF BB, the interaction between Apt/MBA−Au NPs and Au PNNs was weak due to electrostatic repulsion. Their electrostatic interaction increased in the presence of PDGF BB. The ζ-potentials of Apt/MBA−Au NPs 7607

dx.doi.org/10.1021/ac5014207 | Anal. Chem. 2014, 86, 7606−7611

Analytical Chemistry

Article

Figure 1. TEM images of (A) Au PNNs, (B) Apt/MBA−Au NPs, and (C) mixture of Au PNNs (1×) and Apt/MBA−Au NPs (150 pM) in the presence of 50 pM PDGF BB. (D) UV−vis absorption spectra of (a) Au PNNs, (b) Apt/MBA−Au NPs, and the mixture of Au PNNs (1×) and Apt/MBA−Au NPs (150 pM) in the (c) absence and (d) presence of 50 pM PDGF BB.

Figure 2. Investigation of parameters affecting Raman spectra (signals) of 4-MBA. (A) Effect of the concentrations of 4-MBA. (B) Effect of three SERS substrates on SERS spectra of 4-MBA (2 μM). (C) Effects of the concentrations of Apt/MBA−Au NPs on the Raman intensity of 4-MBA at 1580 cm−1. Phosphate solutions (10 mM, pH 7.0) were used in panels A and B. Phosphate solutions (10 mM, pH 7.0) containing 50 pM PDGF BB and 1× Au PNNs, different concentrations of 4-MBA, Apt (2 μM), and Au NPs (15 nM) were used to prepare Apt/MBA−Au NPs in panel A. Thirteen nanometer Au NPs (15 nM), Apt/MBA Au NPs (15 nM), and Au PNNs (1×) were used separately in panel B. Apt (2 μM), 4-MBA (2 μM), and Au NPs (15 nM) were used to prepare Apt/MBA−Au NPs in panel C.

and Au PNNs were −22.3 and −48.6 mV, respectively. In the presence of PDGF BB (50 pM), the ζ-potential of a solution containing Apt/MBA−Au NPs and Au PNNs was 0.3 mV. Optimization of Sensing Conditions. With respect to the specific interaction between Apt and PDGFs and stability of Apt/

MBA−Au NPs and Au PNNs, the experiments were conducted in PBS solution containing 137 mM NaCl at pH 7.0. The two nanomaterials were stable at such a high concentration of NaCl, mainly because they were stabilized separately with the Apt/MBA and SDS. The surface densities of Apt and 4-MBA on the surfaces 7608

dx.doi.org/10.1021/ac5014207 | Anal. Chem. 2014, 86, 7606−7611

Analytical Chemistry

Article

Figure 3. (A) Sensitivity and (B) selectivity of the SERS assay for PDGF BB. Concentrations: 50 pM for PDGF AA, AB, and BB and 1 nM for the other interferences. No PDGFs and interference molecules were presented in the solution as a control. Phosphate solutions (10 mM, pH 7.0) containing 150 pM Apt/MBA−Au NPs and 1× Au PNNs were used to prepare the samples.

concentration of Au PNNs on the Raman signal of 4-MBA was investigated in the presence 50 pM PDGF BB and 150 pM Apt/ MBA−Au NPs as shown in Supporting Information Figure S2, revealing that the SERS signal is proportional to the concentration of Au PNNs up to 1×. Further increases in the concentration of Au PNNs led to less amounts of Apt/MBA−Au NPs per Au PNNs, leading to reduced Raman signals of 4-MBA. Sensitivity, Selectivity, and Reproducibility. Figure 3A shows the SERS signal of 4-MBA at 1580 cm−1 increased upon the increasing concentration of PDGF BB ranging from 1 to 50 pM (R2 = 0.98). This approach provided an LOD at a signal-tonoise (S/N) ratio 3 of 0.5 pM for PDGF BB. Similarly, this approach allowed detection of PDGFs AA and AB down to 0.5 and 0.6 pM, respectively. At such a low concentration of PDGF BB, the SERS signals from PDGF BB were not detected. The high sensitivity is due to the high binding affinity between PDGF and the Apt and a great EF value provided by Au PNNs. The dissociation constant between the free Apt and PDGF BB is about 100 pM, and the binding affinity increases when the Apt molecules are conjugated to Au NPs because of the multivalent binding effect.29 The EF value was calculated according to the equation EF = (ISERSCNR)/(INRCSERS), in which ISERS and INR are the intensities at 1580 cm−1 obtained through SERS and normal Raman scattering measurements, respectively. CSERS and CNR are the concentrations of 4-MBA used for the SERS and normal Raman scattering measurements, respectively. The EF value of this system in the presence of 50 pM PDGF BB was estimated to be 1.3 × 107 by assuming that 4-MBA was all modified on Apt/ MBA−Au NPs. This value was between the EF values of 13 nm Au NPs (1.9 × 106) and Au PNNs (5.6 × 109).16 In addition to a lower EF value, SERS using Apt/MBA−Au NPs alone for the detection of PDGF suffered from more serious matrix interference. Although Au PNNs provide a great EF value, they were not used for preparation of Apt−Au PNNs, mainly because of a difficulty of conjugation due to the existence of SDS molecules on their surfaces that stabilized Au PNNs. Au PNNs not only provided higher signals but also reduced the amounts of expensive Apt/MBA−Au NPs used. Control experiments were carried out to test the specificity of the SERS assay for PDGF (50 pM) over potential interferences, including GSH, cysteine, ATP, hemoglobin, myoglobin, cytochrome c, tripsinogen, and lysozyme (each 1 nM) as shown in Figure 3B. The assay was very selective to PDGFs over the tested interferences, mainly because of high specificity

of Au NPs are important factors in determining the sensitivity for PDGFs. At low density of Apt, Apt tended to form a flat structure that could not interact with PDGFs strongly and Apt−Au NPs were unstable against salt. On the other hand, if the surface density of Apt was too high, its interaction with PDGF BB decreased as a result of steric repulsion. In addition, the surface density for 4-MBA was low, leading to low SERS signals of 4MBA. Supporting Information Figure S1 shows the effect of the concentration on the Apt/MBA−Au NPs. Owing to a high SERS signal and low SERS background, the peak at 1580 cm−1 (aromatic ring vibration of 4-MBA) was used for quantitation.28 The Raman signal of 4-MBA achieved the maximum at 2 μM of Apt. Further increases in the concentration of Apt did not improve the sensitivity, mainly because the concentration of Apt on Au NPs reached saturation at 2 μM.26 With respect to stability and specificity, Apt at 2 μM was selected. Under this condition, 4MBA solutions at various concentrations (0.1−5 μM) were tested. Figure 2A displays the Raman signal reached saturation at about 2 μM of 4-MBA. When the concentration of 4-MBA was further increased, the SERS signal only slightly increased, but the background SERS signals increased. Thus, Apt (2 μM), 4-MBA (2 μM), and 13 nm Au NPs (15 nM) were used to prepare Apt/ MBA−Au NPs. Figure 2B shows slightly different SERS spectra of Apt/MBA−Au NPs, and of 4-MBA when separately using Au NPs and Au PNNs as substrates, in which the concentration of 4MBA used in each case was 2 μM. The Raman peaks at 1130 and 1360 cm−1 (CH bond bending and COO− vibration) are more apparent when using Au PNNs as the substrate, indicating 4MBA was more flat oriented to their surfaces.28 This revealed that less Au−S bonds formed between 4-MBA and Au PNNs, mainly because the steric effect and repulsion caused by the SDS. This explained that Au PNNs relative to Au NPs having a greater EF factor did not provide greater SERS signals of 4-MBA when it was directly attached to Au PNNs. The Apt/MBA−Au NPs solutions at various concentrations (50−300 pM) were investigated for the detection of 50 pM PDGF BB, revealing an optimal concentration of 150 pM as shown in Figure 2C. Upon increasing the concentration of Apt/ MBA−Au NPs over a range from 50 to 150 pM, the SERS signals increased as a result of more MBA molecules on the aggregates. When the concentration of Apt/MBA−Au NPs was further increased, each Apt/MBA−Au NP interacted with less PDGF molecules, leading to decreases in their interactions with Au PNNs and thus decreased SERS signal. The effect of the 7609

dx.doi.org/10.1021/ac5014207 | Anal. Chem. 2014, 86, 7606−7611

Analytical Chemistry

Article

Table 1. Comparison of Various Optical Assays for PDGFs

a

methods

probes

LOD

linear range

ref

colorimetric colorimetric fluorescence FRET luminescence photoluminescence chemiluminescence scattering SERS colorimetric

Apt−Au NPs Apt and Au NPs Apt−diamond and Cy5 Apt−Ag NPs Apt−Ru(II) complexes Apt−Au NPs, PDGF−Au NDs Apt−Au NPs, luminol Apt−Ag NPs Au PNNs, Apt/MBA−Au NPs Apt−microbeadsa

0.3 nM 6 nM 4 pM 27 pM 0.8 pM 80 pM 60 pM