ARTICLE pubs.acs.org/Langmuir
Binary Optical Encoding Strategy for Multiplex Assay Baocheng Tang,†,‡,§ Xiangwei Zhao,†,§ Yuanjin Zhao,† Wendong Zhang,† Qiran Wang,† Lingfang Kong,† and Zhongze Gu*,†,‡ †
State Key Laboratory of Bioelectronics, School of Biological Science and Medical Engineering, Southeast University, Nanjing 210096, China ‡ Laboratory of Environment and Biosafety, Research Institute of Southeast University in Suzhou, Dushu Lake Higher Education Town, Suzhou 215123, China
bS Supporting Information ABSTRACT: A binary optical encoding strategy is proposed to meet the increasing requirements of multiplex bioassays. As illustrated in fluorescence immunodetection of multiplex antigen molecules, photonic crystal beads (PCBs) and quantum dots (QDs) can be used as biomolecular microcarriers and fluorescence labels, respectively. The categories of antigens were deciphered by the binary combination of optical spectra of PCBs and QDs as independent encoding elements. The number of categories that could be detected was theoretically m n, where m and n represent the number of encoding PCBs and QDs, respectively. In addition, the concentrations of the antigens were determined by the fluorescence signals of the QDs. Results of sensitivity analysis indicate that a low-level detection of 58 pg/mL was achieved. Because of the special nanostructures of these two encoding elements, the binary encoding strategy demonstrated its superiority and practicability when compared with single PCB or QD encoding. This supports potential application in multiplex bioassays.
’ INTRODUCTION Multiplex bioassays are of great value for diagnosis, gene expression, and drug screening because of their ability simultaneously to perform multiple reactions in one tube.16 The key technique in multiplex assays is the encoding strategy used to identify different reactions; the most commonly used encoding elements are fluorescent dyes or quantum dots (QDs).711 Fluorescence-encoded microspheres have already been applied to genotyping, proteomics, and diagnostics by the Luminex Corporation.12,13 However, fluorescence from the as-encoded microcarriers will interfere with that from biomolecular labels, and their instability greatly decreases the assay accuracy and sensitivity. Therefore, nonfluorescence encoding is a good alternative. For example, the reflection peak of a photonic crystal (PC) as an encoding element1417 did not suffer from interference, bleaching, quenching, or chemical susceptibility. In addition, the structure of the PCs can provide a higher surfaceto-volume ratio because they are derived from the assembly of colloidal nanoparticles.1822 All of these properties make PCs a novel encoding element suitable for high sensitivity and multiplex bioassays. However, their encoding capacity is limited because the reflection peaks, whose full width at half-maximum (fwhm) is approximately 30 nm, are usually located in the range of visible light (from 400 to 700 nm). Attempts were made via layer-by-layer (LbL) assembly of QDs on the surface of PC beads r 2011 American Chemical Society
(PCBs) to generate QD-coated silica PCB carriers.23,24 In this way, hundreds of codes were obtained by the combination of fluorescence intensity and spectra of QDs and reflection spectra of PCs. However, this encoding method still has the problem of fluorescence interference because fluorescence acted both as a microcarrier encoding element and as a molecular label. It also suffers from the instability of LbL; therefore, we introduce here a binary encoding strategy to achieve a large encoding number utilizing only the spectra of QDs and PCs while eliminating the instability and interference problems. Taking a sandwich immunoassay as an example, the principle of binary optical encoding is shown in Figure 1a. In a multiplex assay of a mixture of antigens, capture antibodies are immobilized on PCB microcarriers, and hence their categories are encoded by the PC reflection spectra (code A). Detection antibodies are labeled with QDs and hence their categories are encoded by QD fluorescence spectra (code B). Because the formation of sandwich immune complexes is specific between the antigens and the capture/detection antibodies, the categories of the antigen are identified by the binary code pair A and B. In addition, the concentration of the antigen can be detected by the fluorescence Received: June 11, 2011 Revised: July 28, 2011 Published: August 08, 2011 11722
dx.doi.org/10.1021/la202202j | Langmuir 2011, 27, 11722–11728
Langmuir
ARTICLE
Figure 1. (a) Schematic representation of the antigenantibody complex (not to scale). Binary optical encoding strategy schemes using PCBs and QDs based on the theory of binary probes. PCBs code A are modified with capture antibody as the capture probe, and QDs code B are modified with detection antibody as the reporter probe. The capture probe and the reporter probe bind specifically to a complementary target antigen to form the antigenantibody complex in solution. (b) Schematic diagram of the experimental method for multiplex assay using silica PCBs and CdSe/ZnS QDs (details are given in the Experimental Section).
signal from the QD label of the detection antibodies as in conventional fluorescence immunodetection. In this case, QDs are only conjugated with detection antibodies and no fluorescence interferences occur. Because of their high brightness and stability, QDs have been increasingly used as substitutes for fluorescence dyes as biomolecular labels.2527 Hence, their combination with PCs in multiplex assays will provide not only high sensitivity but also large encoding capacity, which will be very valuable for practical applications. In this paper, we describe this binary encoding strategy for usual multiplex immunoassays and exploit the traits of PCs and QDs.
’ EXPERIMENTAL SECTION Materials. Monodisperse colloidal silica particles with diameters ranging from 186 to 315 nm were synthesized by the St€ober method. Poly(tetrafluoroethylene) (PTFE) pipes with inner diameters of 500 μm and a T-junction (P-727), which were used to make a coflow microfluidic device, were purchased from Upchurch Scientific, Oak Harbor, WA. Methyl silicone oil was purchased from Yunuo Chemicals Ltd., China. Poly(dimethylsiloxane KF-96) 10 cSt as the continuous phase of the microfluidic device was obtained from Shin-Etsu Chemical, Japan. CdSe/ZnS water-soluble QDs (quantum yield of 40%), 525 nm QD-tagged goat anti-mouse IgG, 525 nm QD-tagged goat anti-rabbit IgG, 605 nm QD-tagged goat anti-mouse IgG, and 605 nm QD-tagged goat anti-rabbit IgG were provided by Wuhan Jiayuan Quantum Dots Co. Ltd., China. rabbit IgG, mouse IgG, goat anti-mouse IgG, goat antirabbit IgG functional fragments, and rabbit serum samples were obtained from Biodee Biotechnology Co., China. (3-Glycidoxypropyl)trimethoxysilane (GPTMS, 98%)28 and bovine serum albumin (BSA) were obtained from Sigma (St. Louis, MO). Phosphate-buffered saline (PBS, 0.05 M, pH 7.4), phosphate-buffered saline Tween-20 (0.05% Tween-20 in PBS), and phosphate buffer (0.05 M, pH 5.0) were prepared in-house. Milli-Q (Millipore, Bedford, MA) water with ultraviolet sterilization was used throughout the experiment.
Instrumentation. A homemade coflow microfluidic device was used to fabricate PCBs as described by Zhao et al.15Bright-field and fluorescence photographs of the beads were taken with an upright microscope (Olympus BX51) and fluorescence microscope (Olympus IX71), respectively. Reflection and fluorescence spectra of PCBs and QDs were recorded with a fiber optic spectrometer (Ocean Optics, QE65000) coupled with the fluorescence microscope. The microstructures of the PCBs were characterized by scanning electron microscopy (SEM; Zeiss, Ultra Plus) and transmission electron microscopy (TEM; JEOL, JEM-2100). Quantitative Assays and Sensitivity Comparisons. For quantitative assays and sensitivity comparisons, different concentrations of mouse IgG were detected via the binary optical encoding strategy. Silica PCB capture probes (preparation methods as shown in Supporting Information) were incubated in test tubes for 30 min (10 μL/bead) at 37 °C. The nonhybridized mouse IgG was removed by rinsing 5 times with PBS followed by electrophoresis for 15 min (80 V, 10 mA). After rinsing and electrophoresis, QD-tagged goat anti-IgG probe solution was used to incubate for 30 min the silica PCBs that had captured the mouse IgG in the tubes. After antibodyantigen complexation, the nonhybridized QD-tagged goat anti-mouse IgG probes were removed by rinsing 5 times with PBS followed by electrophoresis for 15 min (80 V, 10 mA). During these processes, the test tubes were shaken at 37 °C. After these treatments, all of the silica PCBs were excited at 365 nm with a UV source to measure the fluorescence. The procedure for other assays [fluorescein isothiocyanate (FITC) probes and QD-coated silica PCB carriers, FITC probes and silica PCBs, and FITC probes and glass beads] was similar to the binary optical encoding strategy. The range of concentrations was from 10 pg/mL to 100 μg/mL. Here, the number of replicates of our analyte was 10 and the detection result is also calculated from the zero calibration plus three times the standard deviation. Multiplexed Assays. Three types of silica PCBs (reflection peak positions at 476, 623, and 650 nm) and two types of CdSe/ZnS QDs (fluorescent peak positions at 525 and 605 nm) were selected to demonstrate the multiplexed protein assays. Five groups of analyte samples were designed: group 1 (PBS solution), group 2 (mouse IgG in PBS solution), 11723
dx.doi.org/10.1021/la202202j |Langmuir 2011, 27, 11722–11728
Langmuir
ARTICLE
Figure 2. Schematic diagram of encoding by silica PCBs and CdSe/ZnS QDs. The use of four silica PCBs and two CdSe/ZnS QDs as encoding elements of binary probes could generate eight binary codes. The reflection spectra peaks of PCBs were at 476, 505, 623, and 650 nm from left to right, and the fluorescence emission spectra peaks were at 525 and 605 nm from top to bottom. (Top) Microscope images of four silica PCBs. (Left) Fluorescence micrographs of two CdSe/ZnS QDs bound on silica PCBs. (Bottom) Eight hybrid optical spectra. Solid lines are reflection spectra of the silica PCBs, and dashed lines are fluorescence emission spectra of the CdSe/ZnS QDs. group 3 (rabbit IgG in PBS solution), group 4 (mouse IgG and rabbit IgG in PBS solution), and group 5 (mouse IgG and rabbit IgG in PBS solution). Different silica PCB capture probes were synchronously incubated in the test tubes for 30 min at 37 °C (as shown in Figure 1b). After the silica PCBs captured the antigens (groups 14), they were incubated in reporter probe 1 solution (525 nm QD-tagged goat anti-mouse IgG and 605 nm QD-tagged goat anti-rabbit IgG in PBS) for 30 min at 37 °C, while the silica PCBs of group 5 were incubated in reporter probe 2 solution (605 nm QD-tagged goat anti-mouse IgG and 525 nm QDtagged goat anti-rabbit IgG in PBS) for 30 min at 37 °C. Finally, the reflection spectra of the silica PCBs and fluorescence emission spectra of the CdSe/ZnS QDs were recorded by a fiber optic spectrometerequipped microscope and fluorescence microscope, respectively, for analysis. The combined signals from code A and code B were utilized for encoding and decoding, and the fluorescence of code B was used as the index for quantitative detection.
’ RESULTS AND DISCUSSION Design of Binary Optical Encoding Signals on PCBs. In our binary optical encoding strategy, two different spectrum signals were combined for large-capacity encoding and decoding: the reflection signal from PCBs and the fluorescence signal from QDs. The reflection signal originated from the periodic optical nanostructures of the PCBs;29,30 thus, this structural intrinsic property did not suffer from fading, bleaching, quenching, or chemical instability. Here, the silica PCBs that exhibited a very low fluorescence background apparently served as the multiplex microcarriers. In addition, we found that the fluorescence background of the silica PCBs could be further reduced by sintering them at 700 °C.31 Therefore, when the silica PCBs were used as multiplex microcarriers, there was no background fluorescence to interfere with the reporter signals. By changing the diameters of the colloidal silica particles of the PCBs, we obtained a series of reflection signals in the visible range (from 400 to 700 nm), and nine types of silica PCBs are shown in Figure S1a (Supporting Information). The second level of encoding signal was the
fluorescence signal of the QDs. In our study, CdSe/ZnS QDs were selected as the encoding fluorescent labels of the reporter molecules. The CdSe/ZnS QDs are a type of quasi-zero-dimensional core/shell semiconductor nanomaterial, and the photoluminescent emission of CdSe/ZnS QDs could be simply tuned as a function of core size. In addition, the effect of ZnS shell passivation on electronic structure of QDs, photochemical stability, and antiphotobleaching capacity of CdSe/ZnS QDs could be greatly improved. Comparing the photochemical stability of CdSe/ZnS QDs with FITC on PCBs, we found that the fluorescence intensity of CdSe/ZnS QDs did not decrease significantly after 1 h, whereas exponential decay of the signal could not be avoided in the case of FITC (as shown in Figure S2, Supporting Information); that is, CdSe/ZnS QDs had higher photochemical stability. Furthermore, CdSe/ZnS QDs showed symmetrical and narrower emission spectra. All of the aforementioned properties strongly supported the feasibility of adopting CdSe/ZnS QDs as the encoding signal. Here, four types of fluorescent signals from CdSe/ZnS QDs, whose average fwhm was 32 nm, were used in our research, as shown in Figure S1b (Supporting Information). One merit of the binary encoding method is illustrated by the combination of reflected signals of PCBs and the fluorescent signal of QDs in the sandwich immunoassays. Because they are based on two completely distinct mechanisms, encoding signals from PCBs and QDs could be well separated and thus a large encoding capacity can be achieved, avoiding negative effects on the detection sensitivity. For example, Figure 2 shows that eight types of binary codes could be achieved by using four silica PCBs and two CdSe/ZnS QDs. Furthermore, in theory, more than 50 types of silica PCBs could be generated (between 200 and 1100 nm), and 20 types of CdSe/ZnS QDs could be purchased up to now; hence, the number of binary codes could reach more than a thousand types in a single assay. Effects of PCB Microcarriers and QD Reporter Probes on Multiplex Assays. The silica PCBs also acted as microcarriers of biomolecules in multiplex assays in the binary optical encoding strategy. The structure of silica PCBs was the most novel 11724
dx.doi.org/10.1021/la202202j |Langmuir 2011, 27, 11722–11728
Langmuir
ARTICLE
Figure 3. SEM images of the inside of PCBs with no probes after hybridization for detection by QD-encoded reporter probes: (a) high-magnification image of QD-encoded reporter probes on colloidal silica particles inside PCBs before rinsing and electrophoresis washing, (b) high-magnification image of QD-encoded reporter probes on colloidal silica particles inside PCBs after rinsing five times by PBS and electrophoresis washing for 15 min (80 V, 10 mA). Silica PCBs were crushed to observe the inside of the silica PCBs.
characteristic that influenced the sensitivity of detection. Silica PCBs with a diameter of 200 μm were generated by a microfluidic system (preparation methods as shown in Supporting Information). By changing the diameter, D, of the colloidal silica particles, a series of silica PCBs with different reflection spectra could be obtained. SEM and TEM images in Figure S3 (Supporting Information) show that silica PCBs have an ordered nanoporous structure. Three types of pores exist in the silica PCBs, with the smallest having a diameter of 0.15D.32 The diameter of monodispersed colloidal silica particles used to generate encoding microcarriers usually ranged from 200 to 300 nm, and so the smallest pores of the silica PCBs had approximate diameters from 30 to 45 nm. The nanopores were interconnected and extended to the interior of the silica PCBs, which provided channels for biomolecules diffusing onto the surface of the colloidal silica particles for immunoreactions. The structure of the PCBs could improve the biomolecules’ immobilization and reaction because of the higher surface areas, and thus lead to highly sensitive multiplex assays. In addition, the signals of CdSe/ZnS QDs were also used to estimate quantitatively the amount of target antigens. The CdSe/ZnS QDs possess higher quantum yields, leading to a high brightness and large concentration-dependent intensity gradient, which might lead to a stronger detection signal. Although the sensitivity of the multiplex assay could be improved because of more protein molecules diffusing to the inner space of the silica PCBs, nonhybridized molecules might also be nonspecifically adsorbed inside the porous structure of the silica PCBs. These nonhybridized molecules could produce a false fluorescence signal and thus influence the reliability of detection; therefore, it was essential to ensure that the nonhybridized protein molecules were completely removed. Here, we eliminated the nonhybridized molecules by rinsing and electrophoresis washing. We prepared silica PCBs with no capture probe and CdSe/ZnS QD-tagged goat anti-mouse IgG reporter probe to detect mouse IgG to demonstrate the reliability of our approach. The nonspecific fluorescence intensity obviously weakened upon rinsing the PCBs five times with PBS, but there was still a residual fluorescence noise on silica PCBs. The residual noise could be eliminated by electrophoresis. Obviously shown in the figure is that the residual nonspecific fluorescence signal on the silica PCBs was eliminated completely when these were subjected to more than 14 min of electrophoresis treatment (as shown in Figure S4, Supporting Information). SEM images of PCBs before and after electrophoresis are compared in Figure 3,
Figure 4. Relationship between fluorescence intensity and concentrations of target antigens detected with QD as label/PCB as carrier, FITC as label/PCB as carrier, FITC as label/QD-coated silica PCB as carrier, and FITC as label/glass bead as carrier. The intensity data derived from different carriers were subtracted from the negative control of analyte (including the nonspecific adsorption). Error bars represent standard deviations. As the detection concentrations are in a wide range, the relationship is not linear.
panels a and b, respectively; these results also indicate that the nonhybridized molecules could be removed by rinsing and electrophoresis. Therefore, we chose the approach of rinsing five times followed by 15 min of electrophoresis treatment (80 V, 10 mA) to eliminate the nonhybridized molecules in our experiment. Quantitative Assays and Sensitivity Comparison. The sensitivity of the binary optical encoding strategy could be affected by the characteristics of the silica PCB carriers, reporter probe labels (CdSe/ZnS QDs), antibodies (monoclonal or polyclonal), and detectors.33 We explored the effect on the sensitivity of silica PCBs as encoding carriers and CdSe/ZnS QDs as encoding fluorescent labels. An experiment was performed to detect mouse IgG via the binary optical encoding strategy (QD as label/silica PCB as carrier), FITC as label/QDcoated silica PCB as carrier, FITC as label/silica PCB as carrier, and FITC as label/glass bead as carrier (Figure 4). The results clearly indicate that the fluorescence signal intensity of the first 11725
dx.doi.org/10.1021/la202202j |Langmuir 2011, 27, 11722–11728
Langmuir
ARTICLE
Figure 5. Multiplexed assays by silica PCBs and CdSe/ZnS QDs based on the binary optical encoding strategy. (Top) Microscopic images of three silica PCBs modified with different capture antibody functional fragments as capture probes. The reflection peaks of three silica PCBs were 476, 623, and 650 nm from left to right. The silica PCBs with a reflection peak at 476 nm with no probe was used as a control group. The silica PCBs at 623 nm were modified with probes 1 and 2, and the silica PCBs at 623 nm were similar. Probe 1 represents the goat anti-mouse IgG functional fragments, and probe 2 represents the goat anti-rabbit IgG functional fragments. (Left) Five groups of analyte samples (from top to bottom): group 1 (PBS solution), group 2 (mouse IgG in PBS solution), group 3 (rabbit IgG in PBS solution), group 4 (mouse IgG and rabbit IgG in PBS solution), and group 5 (mouse IgG and rabbit IgG in PBS solution). (Middle) Results of multiplexed assays: yellow solid lines are reflection spectra of silica PCBs, and yellow dashed lines are fluorescence emission spectra of CdSe/ZnS QDs. Fluorescence microscope images of CdSe/ZnS QDs reporter probes on the corresponding PCBs are shown on the top of the corresponding spectrum. The reporter probes of groups 14 were 525 nm QD-tagged goat anti-mouse IgG and 605 nm QDtagged goat anti-rabbit IgG, and the reporter probes of group 5 were 605 nm QD-tagged goat anti-mouse IgG and 525 nm QD-tagged goat antirabbit IgG.
shows a much higher numerical value than the latter three. We could detect 34.4, 0.92, and 27 ng/mL mouse IgG with QDcoated silica PCB, silica PCB, and glass bead carriers with FITClabeled reporter probes, respectively; however, the binary optical encoding strategy was able to detect 58 pg/mL at a signal-tonoise ratio of 3:1. Also, our strategy had good reproducibility (the coefficients of variation of fluorescence intensities among different batches analyzed was as low as 6%). Various factors might contribute to the results. First, the porous structure of the silica PCBs provided more reaction surface area and higher surface-tovolume ratio. This means that more probe molecules could capture analytes and more QD-encoded reporter probes participated in photon absorption and subsequent photon emission on silica PCBs than on glass beads and LbL-assembled QD-coated silica PCB carriers, in which the hybridization happened only on the surface layer. Second, the high blank background fluorescence parameters of the QD-coated silica PCB carriers could greatly decrease the sensitivity. In contrast to the QD-coated silica PCB carriers, the silica PCBs as a type of nonfluorescent encoded microcarrier contributed much lower background fluorescence, leading to lower blank background fluorescence parameters in quantitative assays. Moreover, compared with fluorescence dyes, CdSe/ZnS QDs had higher brightness and quantum yield as fluorescent labels and they are considered good substitutes. In addition, after the experiment of checking rabbit
IgG in real rabbit serum samples detection, we have demonstrated that our results are also acceptable in real serum samples. Multiplexed Assays. Because of wide availability, cheap price, and low cross reaction, mouse IgG and rabbit IgG, as the common detection model for multiplex assay, were chosen to demonstrate the reliability and multiplexing capabilities of our strategy for bioassays. Based on the two analytes, five groups of analyte samples were designed [details are shown in Figure 5 (left) from top to bottom]. The multiplex microcarriers of the capture probes were three silica PCBs carriers with reflection peak positions at 476, 623, and 650 nm (as shown at the top of Figure 5). The silica PCBs with the reflection peak position at 476 nm, which did not carry a capture antigen, was used as a control; the second one at 623 nm was modified with two types of capture antibodies, probe 1 (goat anti-mouse IgG) and probe 2 (goat anti-rabbit IgG), as two types of capture probes, and the third one at 650 nm was treated similarly. In addition, two types of CdSe/ZnS QDs with fluorescent peak positions at 525 and 605 nm were modified with goat anti-mouse IgG and goat antirabbit IgG, respectively, and used as the multiplex reporter probes. Because of the specific binding among capture probes, corresponding target, and reporter probes, silica PCB capture probes could capture corresponding antigens in the target samples, and then corresponding QD reporter probes could bind with the antigens that have been captured. By analyzing the 11726
dx.doi.org/10.1021/la202202j |Langmuir 2011, 27, 11722–11728
Langmuir
ARTICLE
reflection signal and the fluorescence signal on silica PCBs, decoding and quantitative analysis of the antigens could be achieved. The results of the multiplex assays were shown in Figure 5. With hybridizing one type of silica PCB (silica PCB at 623 or 650 nm) and two types of QD (CdSe/ZnS QD at 525 or 605 nm), the detection of two target analytes was accomplished at one time [as shown for group 4 (17, 18) or (19,20)]. Similarly, by using one type of QD (CdSe/ZnS QD at 525 or 605 nm) and two types of silica PCB (silica PCB at 623 or 650 nm), we have also achieved the detection of two target analytes at one time [as shown for group 4 (17, 19) or (18, 20)]. In addition, when we exchanged the QD labels between two types of reporter probes, that is, changing 525 nm QD-tagged goat anti-mouse IgG and 605 nm QD-tagged goat anti-rabbit IgG to 605 nm QD-tagged goat anti-mouse IgG and 525 nm QD-tagged goat anti-rabbit IgG, we achieved four types of different detection results [as shown for group 5 (22, 23, 24, 25)]. In other words, just by using another two types of binary probes, another two antigens would also be detected. Therefore, four types of binary optical codes were achieved and the detection of four target analytes could be accomplished at one time by hybridizing two silica PCB codes and two QD codes. More binary optical codes would be achieved for the detection of a large number of analytes by increasing the number of encoded PCBs and QDs. In theory, the number of binary optical codes (N) is based on the number of PCB codes (m) and QD codes (n), and is given by the following formula N ¼ mn
ð1Þ
This type of binary encoding strategy is especially effective and practical when encoding capacity is a consideration because of cost, sensitivity, and decoding accuracy. In this case, although a large number of codes was demonstrated by Nie and co-workers,34 the hybrid encoding method based on different QDs still retains some practical problems in applications, such as signal interference and restriction of the encoding capacity because of limitation to the visible spectrum region for convenient application. However, by combining two different multiplex signals of PCBs and QDs, we will achieve large capacity, high sensitivity, and high selectivity bioassay. In addition, it is worth noting that fluorescence signals resulting from nonspecific adsorption on the microcarriers were almost nonexistent and that hybridization selectivity was improved in our strategy. From the control microcarriers and the other control group [as shown in Figure 5 (16, 8, 1012, 14, and 16)], it is clear that the nonhybridized molecules could be eliminated by rinsing and electrophoresis washing. Such results imply that no other fluorescent signals affected the reporter signals in our experiment. In conclusion, we can generate large-capacity codes by hybridizing the signals from silica PCBs and CdSe/ZnS QDs in binary probes to achieve flexible encoding for reliable and sensitive multiplex assays.
’ CONCLUSION In summary, we have developed a new strategy to achieve largecapacity encoding by hybridizing the encoding signal of binary probes from PCBs and QDs for sensitive multiplex assays. Hybridization selectivity was improved by using binary probes labeled with PCBs and QDs. We chose silica PCBs and CdSe/ZnS QDs as two encoding elements, which had symmetrical and narrow spectra and achieved large-capacity binary encoding. Silica PCB multiplex microcarriers had low background fluorescence
and only the fluorescent signal from CdSe/ZnS QDs was detected for each of the analyte analyses, so there was no background fluorescence and interference signal during detection; thus, the reliability and sensitivity of the assay was improved. The high surface-to-volume ratio of silica PCBs and high brightness of CdSe/ZnS guarantee a highly sensitive assay. The nonspecific fluorescence signal could be effectively eliminated by rinsing followed by electrophoresis. Results of sensitivity analysis indicate that our strategy can achieve a highly sensitive assay and we could achieve detection of 58 pg/mL. Protein multiplex assay analysis showed flexible and large-capacity encoding with high reliability. Our strategy could also be modified with appropriate probes to detect other biological molecules, such as DNA, which illustrates the flexibility and feasibility in potential applications.
’ ASSOCIATED CONTENT
bS
Supporting Information. Additional text, describing generation of silica PCBs, preparation of PCB capture probes and QD-encoded reporter probes, and preparation of QD-coated silica PCB carriers, and four figures, showing reflection spectra of nine silica PCBs and fluorescence emission spectra of four CdSe/ ZnS QDs, comparison plots of photochemical stability between FITC and CdSe/ZnS QDs on silica PCBs, SEM and TEM images of silica PCBs, and relationship between electrophoresis time and fluorescence intensity of residual nonhybridized QDencoded reporter probes. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail
[email protected]. Author Contributions §
These authors contributed equally to this paper.
’ ACKNOWLEDGMENT We are grateful for the support of 333 Talent Project Foundation and the Qing Lan Project of Jiangsu Province, Jiangsu Science and Technology Department (Grants BE2009148 and Bk2008318), and the National Science Foundation of China (Grants 50925309 and 21073033), and Y.Z. thanks the Scientific Research Foundation of Southeast University (Grant Seucx201104). ’ REFERENCES (1) Walt, D. R. Science 2000, 287, 451. (2) Meza, M. B. Drug Discovery Today 2000, 1, 38. (3) Braeckmans, K.; De Smedt, S. C.; Leblans, M.; Pauwels, R.; Demeester, J. Nat. Rev. Drug Discovery 2002, 1, 447. (4) Nam, J. M.; Thaxton, C. S.; Mirkin, C. A. Science 2003, 301, 1884. (5) Rissinand, D. M.; Walt, D. R. Anal. Chim. Acta 2006, 564, 34. (6) Schwartz, M. P.; Alvarez, S. D.; Sailor, M. J. Anal. Chem. 2007, 79, 327. (7) Traul, M.; Battersby, B. J. Adv. Mater. 2001, 13, 975. (8) Cunin, F.; Schmedake, T.; Link, J.; Li, Y.; Koh, J.; Bhatia, S.; Sailor, M. J. Nat. Mater. 2002, 1, 39. (9) Alivisatos, A. P. Nat. Biotechnol. 2004, 22, 47. (10) Leatherdale, C. A.; Woo, W. K.; Mikulec, F. V.; Bawendi, M. G. J. Phys. Chem. B 2002, 106, 7619. (11) Bruchez, M., Jr.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A. P. Science 1998, 281, 2013. 11727
dx.doi.org/10.1021/la202202j |Langmuir 2011, 27, 11722–11728
Langmuir
ARTICLE
(12) Gunderson, K. L.; Kruglyak, S.; Graige, M. S.; Garcia, F.; Kermani, B. G.; Zhao, C.; Che, D.; Dickinson, T.; Wickham, E.; Bierle, J.; Doucet, D.; Milewski, M.; Yang, R.; Siegmund, C.; Haas, J.; Zhou, L.; Oliphant, A.; Fan, J. B.; Barnard, S.; Chee, M. S. Genome Res. 2004, 14, 870. (13) Vignali, D. A. A. J. Immunol. Methods 2000, 243, 243. (14) Zhao, X. W.; Cao, Y.; Ito, F.; Chen, H. H.; Nagai, K. J.; Zhao, Y. H.; Gu, Z. Z. Angew. Chem., Int. Ed. 2006, 45, 6835. (15) Zhao, Y. J.; Zhao, X. W.; Sun, C.; Li, J.; Zhu, R.; Gu, Z. Z. Anal. Chem. 2008, 80, 1598. (16) Zhao, Y. J.; Zhao, X. W.; Hu, J.; Xu, M.; Zhao, W. J.; Sun, L. G.; Zhu, C.; Xu, H.; Gu, Z. Z. Adv. Mater. 2009, 21, 569. (17) Zhao, Y. J.; Zhao, X. W.; Tang, B. C.; Xu, W. Y.; Gu, Z. Z Langmuir 2010, 26, 6111. (18) Umbanhowar, P. B.; Prasad, V.; Weitz, D. A. Langmuir 2000, 16, 347. (19) Kim, S. H.; Lee, S. Y.; Yi, G. R.; Pine, D. J.; Yang, S. M. J. Am. Chem. Soc. 2006, 128, 10897. (20) Kim, S. H.; Jeon, S. J.; Yi, G. R.; Heo, C. J.; Choi, J. H.; Yang, S. M. Adv. Mater. 2008, 20, 1649. (21) Nisisako, T.; Toriib, T. Lab Chip 2008, 8, 287. (22) Park, J. I.; Nie, Z. H.; Kumachev, A.; Abdelrahman, A. I.; Binks, B. P.; Stone, H. A.; Kumacheva, E. Angew. Chem., Int. Ed. 2009, 48, 5300. (23) Li, J.; Zhao, X. W.; Zhao, Y. J.; Gu, Z. Z. Chem. Commun. 2009, 2329. (24) Li, J.; Zhao, X. W.; Zhao, Y. J.; Hu, J.; Xu, M.; Gu, Z. Z. J. Mater. Chem. 2009, 19, 6492. (25) Dabbousi, B. O.; Viejo, R. J.; Mikulec, F. V.; Heine, J. R.; Mattoussi, H.; Ober, R.; Jensen, K. F.; Bawendi, M. G. J. Phys. Chem. B 1997, 101, 9463. (26) Hu, M.; Yan, J.; He, Y.; Lu, H. T.; Weng, L. X.; Song, S. P.; Fan, C. H.; Wang, L. H. ACS Nano 2010, 4, 488. (27) Song, S. P.; Qin, Y.; He, Y.; Huang, Q.; Fan, C. H.; Chen, H. Y. Chem. Soc. Rev. 2010, 39, 4234. (28) Yang, Z. J.; Xie, Z. Y.; Liu, H.; Yan, F.; Ju, H. X. Adv. Funct. Mater. 2008, 18, 3991–3998. (29) Yablonovitch, E. Phys. Rev. Lett. 1987, 58, 2059. (30) John, S. Phys. Rev. Lett. 1987, 58, 2486. (31) Zhao, X. W.; Zhao, Y. J.; Hu, J.; Xu, M.; Zhao, W. J.; Gu, Z. Z. J. Nanosci. Nanotechnol. 2010, 10, 588. (32) Gu, Z. Z.; Hayami, S.; Kubo, S.; Meng, Q. B.; Einaga, Y.; Tryk, D. A.; Fujishima, A.; Sato, O. J. Am. Chem. Soc. 2001, 123, 175. (33) Rosi, N. L.; Mirkin, C. A. Chem. Rev. 2005, 105, 1547. (34) Han, M. Y.; Gao, X. H.; Su, J. Z.; Nie, S. M. Nat. Biotechnol. 2001, 19, 631.
11728
dx.doi.org/10.1021/la202202j |Langmuir 2011, 27, 11722–11728
