1. Introduction

The detection of sequence-specific DNA/RNA hybridization at the single-molecule level has been gradually applied to a variety of biological and biomedical applications, such as clinical diagnostics and biosensors [1, 2]. Conventional DNA/RNA detection methods, including northern blotting, microarrays, and quantitative reverse transcription polymerase chain reaction (qRT-PCR) involve time-consuming and expensive processes that require special analytical equipment [3, 4]. Therefore, label-free systems are of interest because they could potentially provide relatively rapid, simple detection of hybridization events without stringency control and repetitive thermal cycling [5, 6]. In addition, to further increase experimental robustness and reduce complexity, single-step assays are highly desirable. Molecular beacons (MBs) are oligonucleotide probes that can rapidly sense the presence of label-free targets in homogeneous solutions. Owing to their excellent specificity to single nucleotides and ability to directly detect unlabeled targets, MBs have been attracting increasing interest in pathogenic detection and genetic mutation detection [7, 8]. However, similar to most other fluorescence techniques, MBs also suffer from rapid photobleaching; thus, prolonged observation of biological processes is restricted. To address this issue, MB probes have been immobilized on plasmonic nanoparticles, such as silver nanoculsters, gold nanoparticles, or quantum dots for metal-enhanced fluorescence and to improve the limit of detection [9–13]. With advances in nanotechnology, surface-enhanced Raman scattering (SERS) has provided new possibilities for various biosensing applications [14]. After excitation of the metallic nanoparticle surface by incident light, a localized electromagnetic field (surface plasmon) that can interact with molecules adsorbed on the metallic surface is generated. However, current SERS assays for the detection of target biomarkers have somedisadvantages, such as time-consuming separation processes and ON-to-OFF systems (negative assays) [15, 16]. For negative assays, it can be difficult to judge whether the reduction in signal is caused by the target being present or by poor assay performance. In order to improve the efficiency of the OFF-to-ON system, it is important to consider the structural stability and flexibility of the MBs [17, 18]. In our study, we reported a positive read-out SERS-based nanobiosensor for the detection of viral mRNA of interest with a turn-on signal switch. This system is composed of reverse-hairpin MBs (RHBs), which could be transformed into a hairpin-shape for specific target RNA and signal switch, and silver-coated gold nanobowls (RHB-GNB@Ag), which provides an order of magnitude of signal enhancement. To prepare efficient RHBs, the stem-loop ratio of RHBs was adjusted. As a model system to validate the performance of RHBs, we designed RHBs with complementary sequences of influenza A virus matrix (M) protein, which increases pathogenesis and is designated as a type-specific antigen [19]. We demonstrated that optimal RHB-GNB@Ag showed OFF-to-ON signal switching in the presence of target. This promising RHB-GNB@Ag sensing technology showed feasibility as a useful Influenza A assay.

2. The operation and detection principles of the SERS-based RHB sensor

As a proof-of concept, we have designed a reverse hairpin beacon probe labeled with Cy3 to detect the mRNA of influenza A matrix protein, which has been studied as an important modulator of viral replication. The operating and detection scheme is shown in Fig. 1. As shown in Fig. 1(a), linearly structured RHBs, having a Raman label at one end, were immobilized onto silver-coated gold nanobowls (generating RHB-GNB@Ag) via a metal-thiol bond. The signal probe (SP) was designed with a hairpin structure in order to produce a SERS signal upon binding targets; then, the loop was closed to bring to Raman label to the surface of silver-coated gold nanobowls (GNB@Ag). A flag probe (FP) could hybridize to the SP in a partial duplex sequence-specific manner. RHBs positioned the label away from the metallic surfaces. In this open configuration, RHB-GNB@Ag exhibited low SERS intensity (OFF state) as the SERS enhancement decreased. Upon exposure to a target nucleic acid, the flag probe left the surface following a single-displacement process. The target first bound to the target complementary sequence region in the FP and then began to displace the FP from the SP through a branch migration process. Figure 1(b) shows the sequence design of the RHBs. The buffer sequence in the SP was composed of seven cytosines, which could tightly hybridize the SP and FP with a melting temperature (Tm) of 62.1°C in Tris-HCl solution (12 mM Na⁺, 0.8 mM Mg⁺⁺); in contrast, the Tm of RHB without buffer sequence was decreased to 37.3°C. These data suggested that the buffer sequence prevented RHBs from spontaneously realizing a hairpin structured without target molecules. The Tm of the hairpin structured SP screened in this study was over than 60.8°C; therefore, that between the SP and FP was theoretically measured to determine their conformational stability.

 

Fig. 1 (a) Schematic representation of RHB-GNB@Ag detection mechanism. RHB in a reverse-hairpin configuration are immobilized on GNB@Ag substrates. (Left) No SERS signals are observed. (Right) RHB become hairpin configuration and rigid after hybridization with target molecules. (b) The design of the RHB sequence for Influenza A virus detection.

Download Full Size | PDF

3. Screening of RHBs with various stem lengths

As shown in Fig. 2, sequence information for SPs with different stem lengths and FPs. To prepare efficient RHBs, the stem-to-loop ratios of the SPs were adjusted from 0.9 to 1.38 because we assumed that the sensitivity was related to the free energy of the conformational change. S9, S10, and S11 SPs had stems with 9, 10, and 11 bp, respectively. The Tms were theoretically calculated as 60.8, 63.2, and 65°C. Different SPs were investigated for their ability to turn off the signal of the RHBs based on fluorescence analysis. Figure 3(a) shows the fluorescence signal from the RHBs in their reverse-hairpin state (open state) and hairpin state (closed state). With the S9 SP, a large fluorescence signal difference was observed between the open and closed states, indicating that the S9 SP efficiently altered the conformation in the presence of specific stimuli. This result could be explained by the observation that the free energy of the 11S signal probe tended to fold into a stem-loop structure. We next investigated whether the stem-to-loop ratio affected the dissociation of the FP from the SP. The fluorescence spectra in Fig. 3(b) show the fluorescence intensities from the SPs after 0.5 h of incubation with targets ranging from 0 to 100 nM at 37°C. In the presence of 100 nM target, the fluorescence intensity of RHBs was significantly decreased, indicating that the self-hybridization enabled the formation of the stem-loop structure by triggering the strand displacement reaction and releasing the FP. As shown in Fig. 3(c), all DNA strands formed the required DNA structure, as demonstrated by gel electrophoresis. When the target was added, it bound to the FP, and the SP changed its conformation to the stem-loop state. Lane 2 shows the SPs in the reverse-hairpin and hairpin states, and lane 3 shows the FP. RHBs partially hybridized between the SP and FP are presented in lane 4, and lane 7 shows RHB-bound target molecules.

 

Fig. 2 (a) Table of RHB sequence information consisting of signal probe (SP) and flag probe (FP) complementary influenza A virus mRNA. (b) The stem-loop configuration of the SP with various stem-to-loop ratios.

Download Full Size | PDF

 

Fig. 3 Structural flexibility of reverse hairpin beacons (RHBs) using different stem base lengths of 9, 10, and 11 bp. a) Fluorescence diagram of two phase RHBs: (i) reverse-hairpin state and (ii) hairpin state. b) Variations in the fluorescence intensity of RHBs over a target concentration profile. c) Electrophoresis images. Lane 1, 50-kb ladder; lane 2, S9 SP; lane 3, FP; lane 4, RHBs; lane 5, target; lane 6, FP + target; lane 7, RHBs + target.

Download Full Size | PDF

4. Sensitivity and selectivity of RHBs

To further asses the sensing ability of RHBs, we analyzed the selectivity and sensitivity of RHBs in Fig. 4(a). First, to estimate the target selectivity of RHBs, we obtained three different types of mimic DNAs from severe acute respiratory syndrome (SARS) and porcine epidemic diarrhea virus (PEDV). The detection performance of RHBs having perfectly matched sequences was significantly reduced compared with that of mismatched target molecules. In particular, the fluorescence intensity of the matched molecules was decreased by 87%. Second, to estimate the target sensitivity of RHBs, the limit of detection (LOD) of RHBs was determined by 0.1-fold serial dilutions of the target concentration in Fig. 4(b). At target molecule concentrations ranging from 0.1 to 10 nM, the fluorescence signal of RHBs gradually decreased. Owing to the excellent functionality of RHBs in the quenching of the fluorescent signal, the LOD of RHBs was determined to be as low as 0.1 nM. Figure 4(c) shows changes in the fluorescence intensity as the RHBs were cycled three times between the closed and open states by injection with the FP. The cycling efficiency did not decrease after three open/close cycles; thus, these RHBs are expected to endure many open/close cycles without a loss in working efficiency. When laser type and wavelength was continuous wave and 633 nm, respectively. Power used in this process was 170 microWatt and beam diameter was 1.4 micrometers. Applying the Semrock laser damage threshold calculator, we could obtain the following results: Laser fluence and filter laser damage threshold used 633 nm is 11043.4 W/cm² and 11898.5 W/cm². LDT rating calculated by 92.81%, therefore, damage is unlikely to occur.

 

Fig. 4 (a) Fluorescence spectra of RHBs over a target virus RNA concentration profile for 1 h. The types of target virus RNA were influenza A, SARS, and PEDV. (b) Fluorescence graph of limit of detection of RHB. (c) Reusability of RHBs. (i), (iii), (v) Reverse-hairpin state; (ii), (iv) hairpin state.

Download Full Size | PDF

5. SERS sensitivity of the RHB target sequences

The SERS enhancement factors on GNB@Ag were estimated by using rhodamines 6G (R6G). 10⁻³ M R6G solution treated GNB substrate and measured R6G Raman signal intensities, and then calculated by following expression: EF = (I_GNB@Ag /I_R6G) × (N_R6G/N_GNB@Ag). N_R6Gis the numbers of R6G molecules and N_GNB@Ag is numbers of GNB@Ag. I_R6G and I_GNB@Ag are the signal intensities of R6G on the surface of GNB@Ag. SERS enhancement factor calculated by 5 x 10⁹. The feasibility of using the RHB-based SERS system is demonstrated in Fig. 5. To investigate the effects of this system on the SERS properties of RHBs, we used gold nanobowls (GNBs) with a diameter of 350 nm and thickness of 50 nm on Si substrates. Figure 5(a) shows representative scanning electron microscope (SEM) images of bare GNBs and GNB@Ag. After silver coating on GNBs, the morphology of GNBs showed more harsh and rough surfaces. Figure 5(b) shows the Raman spectra of 10 mM R6G molecules absorbed on GNB@Ag. The intensity of the Raman spectra was about ~50 × stronger than bare GNBs. The SERS spectrum exhibited typical ring vibrations of R6G at 1364, 1511, and 1649 cm⁻¹. To turn on the nanobiosensor in the SERS-based system, RHBs were immobilized to the SERS-active GNB@Ag through 5′-HS RHBs. Figure 5(c) shows the SERS signal of the influenza A RHB-GNB@Ag in the presence and absence of target molecules. In the presence of 10 nM synthetic target DNA (green curve), the SERS intensity was significantly increased, indicating that hybridization between targets and the FP enabled the formation of the stem-loop structure, thereby moving the SERS dye onto the GNB@Ag surface and turning the SERS signal “ON”. The major peaks at 1197, 1393, and 1590 cm⁻¹ were assigned to Cy3.

 

Fig. 5 (a) SEM image of a gold nanobowl (left) and silver-coated gold nanobowls (right) (SC). Scale bar: 200 nm. (b) SERS spectrum of GNB@Ag obtained from rhodamine 6G solution (c) SERS spectra after incubation of the RHB-GNB@Ag with the blank and or with complementary target synthetic ssDNA.

Download Full Size | PDF

6. Methods

6.1 Materials

Hydrogen tetrachloroaurate hydrate (HAuCl₄·3H₂O), tris(2-carboxyethyl)phosphine hydrochloride (TCEP), potassium cyanide (KCN), aniline, and (3-aminopropyl)trimethoxysilane (APTMS) were purchased from Sigma-Aldrich (St. Louis, MO, USA). N-Methyl-2-pyrrolidone (NMP), methyl alcohol, sulfuric acid, hydrogen peroxide, nitric acid, and hydrochloric acid were purchased from Duksan Pure Chemicals Co. Functionalized DNA oligonucleotides and viral mRNA mimic DNA were purchased from Bioneer Inc. (Daejeon, Korea). Ultrapure deionized water was used for all the syntheses. Rhodamine 6G was obtained from Sigma-Aldrich Korea. All other chemicals and reagents were of analytical grade.

6.2 Characterization of RHBs

The stem-loop configuration was predicted using the two-state folding tool on the DINAMelt server. To assess the ability of RHBs to affect a target molecule, we used a synthetic target oligonucleotide and mismatched molecules. RHBs were incubated with target or nontarget oligonucleotides at various concentrations for 0.5 h at 37°C, followed by measurement of the fluorescence intensity using a hybrid multimode microplate reader. To evaluate conformational changes in RHBs, the structures of RHBs in the reverse-hairpin and hairpin states were confirmed by gel retardation assay. All DNA oligomers were mixed with 6 × HiQ goRed (Genepole, Seoul, Korea), loaded onto 2% agarose gels (w/v), and electrophoresed in Tris-borate-ethylenediaminetetraacetic acid (TBE) buffer at 100 V for 30 min. The retardation of complexes was visualized by a UV lamp using a Gel Doc System.

6.3 Preparation of GNBs

To prepare GNBs, 250-nm-thick polymethylglutarimide (PMGI; PMGI SF5; MicroChem Corp., USA) was first spin-coated on an Si substrate. Next, a 100-nm-thick thermal nanoimprint resist (mr-I 8010R; Micro-resist Technology GmbH, Germany) was spin-coated. The bilayer resist stack was subjected to thermal nanoimprinting at 200°C under a pressure of 30 bar for 180 s with a commercial tool (ANT-4H; Extech, Korea). We used an array of Si pillars with a 600-nm pitch and a pillar diameter of 300 nm (Eulitha AG, Switzerland) as the nanoimprint mold. After nanoimprinting, the residual nanoimprint resist was etched by O₂ plasma, and the PMGI resist was then wet-etched by a commercial developer solution (AZ MIF300; AZ Electronic Materials, USA) for 3 s. Finally, the sample was mounted in a conventional thermal evaporation chamber at an oblique angle with respect to the evaporation source and was continuously rotated during the deposition to form the nanobowl-shaped structure.

6.4 Preparation of GNB@Ag

Prior to coating, the substrates were rinsed with deionized water in three times. In order to passivate, we lacquered unexposed area for the growth silver nanostructures in the nanobowls. Three-electrode system with a Pt wire counter electrode and Ag/AgCl (1 M KCl) reference electrode were used to electrochemical using potentiostat (CompactStat, Ivium). Silver solution was prepared of 20 mM silver salt and regulated pH using Na₂CO₃ to grow silver particles into gold nanobowls. Gold nanobowls were soaked with silver solution, and then silver nanoparticles were plated at −3.0 V versus Ag/AgCl until −200 mC cm⁻² at room temperature and polymer was removed with excess acetone (Duksan, Korea). The morphology of the GNB@Ag was analyzed by scanning electron microscope (JSM-7001F; JEOL, Japan).

6.4 Fabrication of RHB-GNB@Ag

GNB@Ag (0.5 × 0.5 mm²) were incubated in 0.5 mL probe solution containing 4 μM RHBs and 0.5 M NaCl in 10 mM Tris-HCl buffer (0.01 M sodium dodecyl sulfate, 0.01 M phosphate, pH 7.4) for 12 h at room temperature. The substrates then were rinsed with 0.5 mL phosphate-buffered saline containing 10 mM Tris-HCl three times to remove unstable RHBs.

6.5 RHB-GNB@Ag SERS assay procedure

Prior to the Raman analysis, solutions of 1 mM rhodamine 6G were prepared. The fabricated substrates were immersed in the solution for 4 h and dried overnight. Raman spectroscopy measurements were performed with confocal Raman microscopy (LabRAM ARAMIS, Horiba). To obtain high-quality images, a 100 × microscope objective (Nikon, NA = 0.95) and an He-Ne laser (λ = 633 nm) were used. The laser power on the sample was approximately 170 μW, and the integration time was 10 s. A silicon wafer with a Raman band at 520 cm⁻¹ was used as the reference for calibration. LabSpec 5 software (Horiba) was used for spectral and image processing and analysis.

7 Conclusion

In this study, we developed a label-free and OFF-to-ON SERS platform using RHB-GNB@Ag for in vitro viral RNA detection. We experimentally confirmed that a stem-to-loop ratio of less than 1 was efficient for detecting viral RNA. The optimized RHBs showed dramatic changes in fluorescence intensity in the presence of the target molecule, which triggered enhancement of the SERS signal. With efficiently designed RHBs, we detected the target within 30 min of incubation with target synthetic DNA molecules. From this positive-readout-based SERS and simple detection system, we found that RHB-GNB@Ag represented a promising diagnostic tool and may have point-of-care applications.