SERS Molecular Sentinel Nanoprobes for Viral Infection Diagnostics

Abstract

In this paper, we describe a surface-enhanced Raman scattering (SERS)-based detection approach, referred to as “molecular sentinel” (MS) plasmonic nanoprobes, to detect an RNA target related to viral infection. The MS method is essentially a label-free technique incorporating the SERS effect modulation scheme associated with silver nanoparticles and Raman dye-labeled DNA hairpin probes. Hybridization with target sequences opens the hairpin and spatially separates the Raman label from the silver surface thus reducing the SERS signal of the label. Herein, we have developed a MS nanoprobe to detect the human radical S-adenosyl methionine domain containing 2 (RSAD2) RNA target as a model system for method demonstration. The human RSAD2 gene has recently emerged as a novel host-response biomarker for diagnosis of respiratory infections. Our results showed that the RSAD2 MS nanoprobes exhibits high specificity and can detect as low as 1 nM target sequences. With the use of a portable Raman spectrometer and total RNA samples, we have also demonstrated for the first time the potential of the MS nanoprobe technology for detection of host-response RNA biomarkers for infectious disease diagnostics.

1. Introduction

There is a strong need to develop diagnostic technologies that can be used at the point-of-care to detect infectious diseases. A promising approach involves detection of the host response to various pathogens by evaluating changes in gene expression in peripheral blood samples, induced in response to infection [1–4]. These studies utilized laboratory-based cDNA microarray systems to characterize changes in RNA transcript abundance as a response to infectious challenges. Based upon research results, the development of host gene expression-based classification systems for discriminating types of infection is quite promising; thus, there is a strong need to develop practical diagnostic systems for use at point-of-care settings.

As a proof of concept, we sought to develop a point-of-care RNA detection and quantification approach to ultimately increase applicability of this strategy. Thus, we chose a single transcript, the human radical S-adenosyl methionine domain containing 2 (RSAD2) gene, as the test system for use in development of this novel RNA detection method. The RSAD2 gene is involved in antiviral defense and is directly induced by human cytomegalovirus (HCMV). The RSAD2 gene is a component of the group of genes found to accurately classify acute respiratory viral infection in comparison to no infection or bacterial infection [3], and is up-regulated in the nasal epithelium of individuals infected with rhinovirus [5]. The human RSAD2 gene encodes a protein known as Viperin, first identified in HCMV infected fibroblasts [6]. Upon viral infection, the type I interferons (IFNs) are produced and secreted by infected cells to initiate a complex signaling cascade, leading to the induction of hundreds of genes that limit viral infection. RSAD2 has been recognized as one of the most highly induced genes upon interferon stimulation or infection with various viruses, including HCMV, influenza virus, hepatitis C virus (HCV), dengue virus, alphaviruses, and retroviruses such as human immunodeficiency virus (HIV) [7]. The RSAD2 protein (Viperin) was shown to localize to the endoplasmic reticulum (ER) and lipid droplets, where it exerts its antiviral function. It has been postulated that RSAD2 may alter lipid droplet formation or the ability of the viral proteins to localize to this organelle, leading to inhibit the replication of various DNA and RNA viruses [7].

By evaluating changes in host gene expression profiles in response to viral infection, Zaas et al. have developed a robust blood mRNA expression signature that distinguishes individuals with symptomatic acute respiratory infections (ARIs) from uninfected individuals with over 95% accuracy [3]. This “acute respiratory viral” biosignature encompasses 30 transcripts of genes known to be related to the host immune response to viral infection. In particular, RSAD2 was the most highly expressed gene in symptomatic individuals from all three human viral challenge studies with live rhinovirus, respiratory syncytial virus, and influenza A.

In this paper, we describe a plasmonics-based detection approach, referred to as “molecular sentinel” (MS) nanoprobes, to detect human RSAD2 gene sequences. The MS approach developed previously in our laboratory incorporates the surface-enhanced Raman scattering (SERS) effect modulation scheme associated with metal nanoparticles and stem-loop DNA probes tagged with Raman labels [8–10]. The MS approach was first demonstrated for the detection of polymerase chain reaction (PCR)-amplified DNA of human immunodeficiency virus type 1 (HIV-1) [8]. Recently, we have also demonstrated the detection of BRCA1 single-nucleotide polymorphisms (SNPs) in breast cancer using the MS method [9], as well as the multiplex capability of the MS nanoprobes for the detection of multiple breast cancer biomarkers in one sample solution [10]. In this study, we aimed to demonstrate the potential of using the MS nanoprobes to detect a host biomarker for infectious disease diagnostics and point-of-care applications. Furthermore, whereas previous works involved method development using synthetic or amplified samples, this work discusses for the first time the detection of specific mRNA transcripts from total RNA sample extracted from human lymph nodes without amplification steps.

As illustrated in Figure 1 (left), hairpin probes having a Raman label at one end are immobilized onto a metal (e.g. Ag) nanoparticle via a thiol group attached to the other end. The metal nanoparticle is used as a signal-enhancing platform (nano-enhancer) for the SERS signal associated with the label. The Raman enhancement is determined by the plasmonic effect at the metal surface (surface plasmon). According to classical electromagnetic theory, suitable metal nanostructures with sizes on the order of tens of nanometers are able to enhance the intensity of incident electromagnetic radiation [11, 12]. These field enhancements can be quite large (10⁶-to 10⁷-, with up to 10¹⁵-fold enhancement at “hot spots”). The intense localized fields can then interact with molecules at or near the metal surface [13–16], and allow single-molecule detection to be achieved [17–19]. In addition, it has been reported that nanoparticle-conjugated oligonucleotides exhibit remarkably sharp melting profiles when compared to unmodified oligonucleotides [20–22]. By taking advantage of this unusual hybridization property, several nanoparticle-based DNA detection methods with improved sensitivity have been previously reported [23–28].

Figure 1.

Detection scheme of the molecular sentinel (MS) nanoprobes.

In the detection strategy of our SERS MS nanoprobes, we exploit the dependence of SERS enhancement upon the distance between the metallic nanoparticle and the Raman label. Over the last two decades, our laboratory has devoted extensive effort to develop the SERS technique for chemical and biological sensing [29–34]. Theoretical studies of the SERS effect have shown that the SERS enhancement, defined as the “G factor”, falls off as G = [r/(r+d)] ¹² for a single molecule located a distance d from the surface of a metal particle of radius r [17]. Thus, the SERS enhancement decreases significantly with increasing distance, due to the decay of a dipole over the distance (1/d)³ to the fourth power. Since the SERS enhancement field decreases rapidly as distance to the surface increases, a molecule (e.g., the Raman label) must be located within ~10 nm of the metal nanoparticle surface in order to experience the enhanced local field. As shown in Figure 1, in the absence of target molecules (DNA or RNA), the hairpin configuration has the Raman label in close proximity to the nanoparticle (closed state) and exhibits a high SERS signal (Figure 1, left). However, when complementary DNA/RNA targets are recognized by the MS nanoprobes, hybridization occurs and the Raman label is separated away from the metal nanoparticle. As a result, the SERS signal of the Raman label is significantly reduced (open state), indicating target recognition and capture (Figure 1, right).

2. Experimental

2.1. Reagents

Silver nitrate (99.995%) was purchased from Alfa Aesar (Ward Hill, MA). Hydroxylamine hydrochloride, 6-Mercapto-1-hexanol and Tris-HCl buffer (pH 8.0) were purchased from Sigma-Aldrich. All solution was prepared with deionized water (18 MΩ-cm).

2.2. Preparation of silver nanoparticles

The silver nanoparticles were prepared by using hydroxylamine hydrochloride as the reduction agent [35]. Briefly, 10 mL of silver nitrate solution (10⁻² M) were added to 90 mL of a hydroxylamine hydrochloride solution (1.67 × 10⁻³ M) containing 3.3 × 10⁻³ M NaOH under vigorous stirring. The colloidal solutions were then stored at 4 °C and used within a few days.

2.3. Preparation of SERS MS nanoprobes

In this study, the DNA hairpin probe, complementary target and non-complementary negative control oligonucleotides, shown in Table S1 (Supplementary data), were all synthesized by Integrated DNA Technologies (IDT, Coralville, IA). The MS nanoprobes were synthesized as described previously [10]. Briefly, silver nanoparticles were incubated with 0.5 μM thiolated hairpin probes containing 0.5 mM MgCl₂ and allowed to react for few hours at room temperature. The functionalized silver nanoparticles were exposed to 6-mercapto-1-hexanol (MCH) to passivate the silver surface. MCH is a commonly used spacer thiol to displace non-specifically adsorbed DNA molecules and raise the surface-bound DNA off the surface so that the DNA chains are bound solely by the sulfur atom [36]. The MS nanoprobes were then purified three times by centrifugation at 12,000 rpm for 10 min, and then resuspended in 20 mM Tris-HCl buffer (pH 8.0).

2.4. SERS measurements

SERS measurements were performed using a Renishaw InVia confocal Raman microscope. A 50 mW HeNe laser (Coherent, model 106-1) emitting a 632.8 nm was used for excitation. The sample solution was kept in a glass tube mounted on an X-Y-Z translational stage. The light from the laser was passed through a laser line filter, and focused into the sample solution with a 10x microscope objective. The Raman scattered light was collected by the same objective, and filtered with a holographic notch filter to block light due to Rayleigh scatter. An 1800 groove/mm grating was used to provide a spectral resolution of 1 cm⁻¹. Raman scattering was detected by a 1024 × 256 pixel RenCam CCD detector. The SERS spectra were acquired using a 10-sec integration time and processed with WiRE 2.0 software (Renishaw).

For RNA detection, a portable bench-top Raman spectrometer with a 3-mW HeNe laser (Advantage 633, DeltaNu) was used. In this study, we have selected the DeltaNu instrument because its small suitcase size makes it appropriate for point-of-care applications. All spectra reported here were smoothed by the Savitzky-Golay filter to remove high-frequency noise while preserving low-frequency Raman spectra. The fluorescence background was removed using a numerical algorithm developed in our laboratory, which uses a moving window to locally determine the fluorescence background level.

2.5. Quantification of DNA probes loaded on silver nanoparticles

The number of DNA probes attached to a silver nanoparticle was determined by a ligand exchange process described previously [37]. Briefly, MS nanoprobes were incubated with Mercaptoethanol (0.5 M) for at least 20 hours at room temperature to release dye-labeled DNA probes from the nanoparticle surface. The solutions were centrifuged at 12,000 rpm for 10 min to isolate the released DNA probes from nanoparticles. The fluorescence emission of the collected supernatants was then measured using the FLUOstar Omega microplate reader (BMG Labtech, Inc.). The concentrations of the released probes were determined according to a standard curve. The number of oligonucleotides per particle was then determined by dividing the total number of released probes by the number of nanoparticles.

3. Results and Discussion

To demonstrate the potential of the MS nanoprobe technology for detection of respiratory diseases, a MS nanoprobe for the RSAD2 gene was designed (RSAD2-MS). The RSAD2 nanoprobe consisted of a 40-base DNA hairpin probe modified with the Cy3 Raman dye on the 3′ end and a dithiol substituent at the 5′ end. The underlined sequences in Table S1 (Supplementary data) represent the complementary arm sequences which form a stem-loop structure. The 23-base loop region between the two complementary “arms” was designed to be complementary to a portion of the RSAD2 gene sequence. Figures 2a and 2b show the transmission emission microscopy (TEM) images of bare Ag nanoparticles and RSAD2-MS nanoprobes, respectively. The hydrodynamic size distribution of the bare nanoparticles shown in Figure S1 (Supplementary data) was measured by using NanoSight NS500 (NanoSight Ldt. Amesbury, UK). The hydrodynamic sizes of the bare nanoparticles and oligonucleotide-functionalized MS nanoprobes were then determined to have mean diameters of 51.6 ± 26.2 and 65.7 ± 19.7 nm, respectively. This size of Ag nanoparticles was chosen as they are close to the optimal size of spherical Ag nanoparticles for SERS, which is around 50 nm [38]. The approximately 14-nm difference between bare nanoparticles and MS nanoprobes was due to coating of the 40-base DNA probes.

Figure 2.

(a) TEM image of the silver nanoparticles. (b) TEM image of the RSAD2-MS nanoprobes. (c) UV-Vis absorption spectra of the bare silver nanoparticles in H₂O (solid line), the RSAD2-MS nanoprobes in 20 mM Tris buffer (dotted line), and the RSAD2-MS nanoprobes in 20 mM Tris buffer containing 2.5 mM MgCl₂ (dashed line).

To determine the number of DNA probes attached to a silver nanoparticle, a ligand exchange process was performed [37]. It was found that the number of DNA probes attached on a nanoparticle was estimated to be ~70, and 80 strands/particle after incubation with 0. 25 mM MgCl₂ for 2 and 20 hours, respectively. By increasing the concentration of MgCl₂ to 0.5 mM, the DNA loading can be increased to ~80 and 90 strands/particle after incubation for 2 and 20 hrs, respectively. Thus, the DNA probe loading can be controlled by varying MgCl₂ concentration and the incubation time. The surface coverage of DNA probes on a nanoparticle is a critical parameter for SERS signal and detection sensitivity. High surface probe density may result in low SERS signal (i.e. affecting stem-loop structure formation) and low hybridization efficiency for target strands. Future studies will be carried out to optimize the surface probe density to improve the detection sensitivity.

The bare Ag nanoparticles and oligonucleotide-functionalized MS nanoprobes were then characterized using ultraviolet-visible (UV-Vis) absorption spectroscopy. As shown in Figure 2c, the MS nanoprobes exhibited an 8 nm red-shift of the absorption peak from 410 to 418 nm indicating that the oligonucleotide probes were conjugated to the Ag nanoparticles. For the formation of hairpin structures, the MS nanoprobes were incubated in a 20 mM Tris-HCl buffer solution containing 2.5 mM MgCl₂. By comparing the absorption spectra shown in Figure 2c, the MS nanoprobes were found to be relatively stable in this Tris/MgCl₂ solution, which was then used as the hybridization solution for the subsequent experiments. Figure 3 shows the specificity of the RSAD2-MS nanoprobes using 100 nM of synthetic DNA as the target molecules. The hybridization was carried out in a glass tube containing 0.1 mL of 20 mM Tris-HCl buffer solution (pH 8.0) and 2.5 mM MgCl₂. The result indicates that the SERS intensity is significantly reduced in the presence of its complementary DNA targets (lower spectrum). On the other hand, the SERS intensity remains high in the absence of DNA (upper spectrum) or the presence of non-complementary DNA (middle spectra), indicating that the MS nanoprobes are in the closed (stem-loop) state. The decreased SERS signal indicates that hybridization with the complementary target DNA molecules opens the stem-loop structure of the probe DNA molecules, thereby separating the SERS dye Cy3 from the silver surface. While storing the MS nanoprobes in a 20 mM Tris-HCl buffer solution at 4°C for over one month, we have found that the MS nanoprobes still maintained their functionality (Figure S2, Supplementary data). Although additional tests for a longer period of time will still need to be conducted in future studies, this result demonstrates the potential of using long-term stored MS nanoprobes for point-of-care applications. To demonstrate that the observed reduction in SERS intensity is due to the plasmonic modulation effect (upon stem-loop opening) and not due to dissociation of the hairpin probes from Ag nanoparticles, we also measured the fluorescence signal of the labels on the MS probes. The fluorescence intensities of the target-incubated MS samples were compared before and after centrifugation (Figure S3, Supplementary data). The supernatant exhibits only ~ 2% of the total fluorescence signal, thus indicating that most of the hairpin probes were still attached to the nanoparticles after hybridization to the complementary targets.

Figure 3.

SERS spectra of the RSAD2-MS nanoprobes in the presence or absence of complementary DNA targets. Upper spectrum: blank (no target DNA present). Middle spectrum: in the presence of 100 nM non-complementary DNA (negative control). Lower spectrum: in the presence of 100 nM complementary target DNA. Five SERS measurements were performed per sample and averaged into a single spectrum.

We performed further studies to demonstrate the possibility for quantitative DNA detection for the RSAD2 gene. In a series of measurements, we incubated the RSAD2-MS nanoprobes with various concentrations of RSAD2 target DNA between 0 and 100 nM. Five SERS measurements were performed per sample and averaged into a single spectrum. Figure 4 illustrates that the reduction of SERS intensity at 1196 cm⁻¹ was greater with increasing RSAD2 DNA target concentration. The result confirms that more Raman labels on the RSAD2-MS nanoprobes were being separated from the silver surface during the hybridization process with increasing concentration of target molecules. By monitoring the SERS intensity of specific peaks, the concentration of DNA target can be estimated indirectly. As shown in the inset of Figure 4, a power fit with a correlation coefficient r of 0.9904 over the DNA target concentration from 1 to 100 nM was obtained. The MS method showed a good reproducibility with an average relative standard deviation (RSD) less than 10%. The results indicate that the MS technique can detect 1 nM and lower target DNA, which is more than 250-times improvement compared to our previous report [8, 9]. At higher target concentrations, SERS signals from samples having over 50 nM target DNA decreased and reached a plateau, which indicated that the probe was fully hybridized and the observed signal was from the background. The general trend in this study indicates the potential of using the MS nanoprobes for quantitative analysis. Together with the multiplex capability of the MS technique shown in our previous study [10], the MS nanoprobes could be a useful diagnostic tool for discriminating types of infection based on the expression profiles of multiple biomarkers.

Figure 4.

Quantitative detection using RSAD2-MS nanoprobes with various concentration of the RSAD2 target DNA from 0 to 100 nM. Five SERS measurements were performed per sample and averaged into a single spectrum.

The target DNA concentrations at both the low end and the high end could be extended by adjusting the total amount of hairpin probes used for the measurements. With the same surface coverage of DNA hairpin probes on nanoparticles, the total amount of hairpin probes could be adjusted by varying the sample volume or nanoprobe concentration. The detection sensitivity for target concentration at the low end could be improved by using smaller amounts of DNA hairpin probes. However, the maximum target concentration that could be quantified would be reduced. Therefore, the dynamic range of the MS nanoprobes could be tuned by simply altering the sample volume or nanoprobe concentration. Thus, by using the above strategy, it is expected that the MS method could be extended to pM or even to sub-pM sensitivity.

To demonstrate the potential of using the MS nanoprobes to detect RSAD2 RNA targets for point-of-care applications, we used a small, portable Raman spectrometer (Advantage 633, DeltaNu) for SERS measurements, and purified RNA samples as the specimen. For the initial “proof-of-concept” test, the total RNA from human lymph node tissue was used for the detection of RSAD2 background expression levels. The RSAD2-MS nanoprobes (0.05 mL) were incubated with human lymph node total RNA (purchased from Applied Biosystems) followed by addition of a 0.05 mL of Tris/MgCl₂ buffer solution (MgCl₂ final concentration: 2.5 mM). The sample solution was allowed to react at room temperature for at least 20 min prior to SERS measurements. As shown in Figure 5, the SERS signal of the RSAD2-MS nanoprobes (solid lines) was relatively reduced in the presence of 0.5 μg (spectrum b) and 1 μg total RNA (spectrum c) compared to the blank sample (spectrum a), and the SERS signal was further reduced in the presence of more RNA sample. To confirm the detection specificity, we have utilized a Cy5.5-labeled MS nanoprobe targeted to the gene sequences of the Influenza A virus (H1N1) nucleocapsid protein (NP) as the control nanoprobe, which is not expected to detect complementary targets in the normal human RNA samples. The results from the control experiments (dotted lines shown in Figure 5) show that the intensity of the major Cy5.5 SERS peaks at 1339 cm⁻¹, 1461 cm⁻¹ and 1625 cm⁻¹ remain high in the presence of normal human total RNAs, confirming the detection specificity using the RSAD2-MS nanoprobes. The current DeltaNu Advantage 633 Raman spectrometer only provides a spectral resolution of 10 cm⁻¹ due to the use of a low dispersion grating. As a result, the SERS spectra measured here were relatively broader than those using the confocal Raman microscope with a spectral resolution of 1 cm⁻¹ (as shown in Figure 3). However, the result shown in Figure 5 demonstrates that the spectral resolution provided by the portable spectrometer is adequate for diagnostics. It is noteworthy that we used a portable Raman spectrometer (small suitcase size) as a reader of the SERS signals. The portable Raman spectrometer has several advantages, such as low cost (~$10,000) and high portability (10-times smaller than the Renishaw confocal Raman microscope, which costs ~$300,000), which make it practical for point-of-care applications. This work also demonstrates that the MS technique can be easily integrated with a small portable Raman spectrometer for point-of-care applications, and the MS technique has the potential for RNA biotarget detection.

Figure 5.

SERS detection of RSAD2 RNA in human lymph node total RNA using RSAD2-MS nanoprobes (solid lines) and control-MS nanoprobes (dotted lines). (a) blank (no RNA present). (b) in the presence of 0.5 μg total RNA. (c) in the presence of 1 μg total RNA.

4. Conclusions

In conclusion, we have demonstrated the feasibility and specificity of using the molecular sentinel technique to detect human DNA and RNA targets, which can be used as a novel host marker in response to viral respiratory infection. We have also shown that the reduction of SERS intensity was dependent on the concentration of target molecules, illustrating the capability to use MS nanoprobes for quantitative analysis. Due to narrow absorption bands and large spectral range, Raman provides great possibilities for multiplexing detection. The multiplex capability, which allows the monitoring of a large number of molecular processes simultaneously, is an important feature in medical diagnostics of a large number of biotargets.

We have demonstrated for the first time the use of the MS nanoprobes to detect specific mRNA transcripts from total RNA sample extracted from human lymph nodes without amplification steps. By using a portable Raman spectrometer for SERS measurements, this MS method shows great potential for point-of-care applications. The results of this study demonstrate that the MS technique can provide a novel diagnostic approach for detecting RNA. The capability to detect and quantify nucleic acid molecules (DNA/RNA) could allow the MS diagnostics technology to play an important role in the diagnosis of infection using host-based transcriptional changes as medical diagnostics biomarkers.

Highlights.

• A novel SERS-based detection method for viral infection diagnostics was developed.

• MS technique possesses high selectivity and multiplexing capability.

• The SERS measurement is simple and can be performed without washing steps.

• The potential for semi-quantitative detection was demonstrated.

• The point-of-care implementation using the MS technique was demonstrated.