1.

Introduction

Detection and tracking of biomolecules is invaluable not only in the field of medical diagnostics but also in biological and biotechnology research, drug development, environmental monitoring, forensic investigations, and biodefense. To this end, researchers of different expertise have sought to develop methodologies/techniques that are sensitive, specific, robust, high throughput (amenable to multiplexing), simple, and cost effective. Presently, most methods used to detect biomolecules that require molecular recognition events (e.g., formation of antibody-antigen or aptamer-analyte complexes) are variations of the “sandwich type” enzyme-linked immunosorbent assays (ELISA) and typically rely on the use of receptors tagged with probe molecules, which may not always be desirable.^{1, 2, 3} Labeling of biomolecules is a time-consuming process that can lead to the loss of biological activity. Other assays make use of nanoparticles and their tendency to form aggregates, which results in color change due to a shift in the plasmon resonance in response to binding events.^{4, 5} While this is a convenient method, it is limited by its susceptibility to changes in the environment, e.g., change in ionic strength or pH, which can also induce aggregation and color change. Another nanoparticle-based technique takes advantage of the ability of metal nanoparticles to quench emission from nearby fluorescent compounds. In this case, the presence of the analyte is detected by the change in the intensity of emission from the fluorescent molecule attached to the receptor.^{6, 7} Direct detection techniques such as mass spectrometry and surface plasmon resonance (SPR) spectroscopy are also widely used, but they are expensive and potentially limited by extensive sample preparation. For example, mass spectrometry requires isolation and purification of the analyte (usually by chromatography).⁸ SPR, which detects binding of the molecule by measuring the change in the refractive index at the surface of the metal functionalized with a conjugated receptor, is subject to nonspecific interactions.

Another technique that is currently under development is based on surface-enhanced Raman scattering (SERS). SERS occurs when a molecule is in close proximity to the metal surface. When combined with the resonance Raman effect, signal enhancements of up ${10}^{14}$ to ${10}^{15}$ have been observed, allowing even single molecule detection⁹ under very favorable conditions. This huge signal enhancement is attributed to enhanced electromagnetic fields within the immediate vicinity of the metal upon excitation of plasmon resonances by photon interaction,¹⁰ and to charge transfer processes between the metal and the adsorbed molecule.¹¹

SERS-based assays,^{3, 12, 13, 14} which require molecular recognition events, have been made possible by labeling either the capture agent or the target molecule itself with a Raman-active molecule. Recently, Neuman ¹⁵ reported SERS-based direct detection of the platelet derived growth factor (PDGF) by monitoring the randomness of the SERS spectra of the PDGF aptamer, following its exposure to the target molecule PDGF. The change in the reproducibility of the pristine anti-PDGF aptamer was thought to be due to change in the conformation of the aptamer as a result of protein binding.

Aptamers are functional single-stranded oligonucleotides (DNA or RNA) generated by the process called systematic evolution of ligands by exponential enrichment (SELEX).¹⁶ They bind to their target molecules selectively and with high affinity by forming secondary structures and shapes (e.g., quadruplex, hairpin loop, and T-junction).^{17, 18} Several aptamers have been developed for the detection of certain biological and chemical threat agents [e.g., ricin,¹⁹ anthrax spores,²⁰ and trinitrotoluene (TNT)²¹]; cancer biomarkers;^{22, 23, 24, 25} HIV-associated proteins;^{26, 27, 28, 29} food-borne pathogens;³⁰ and other biologically important biomolecules such as insulin,^{31, 32} immunoglobulin E (IgE),³³ and thrombin.³⁴ They exhibit protein binding affinities that are comparable to those of corresponding antibodies.^{18, 35, 36, 37, 38} As in the case of antibodies, labeling of the aptamer can sometimes lead to loss of activity resulting from loss of ability to form the secondary structure that is required for binding.

In the present study, we exploit the sensitivity offered by SERS in the development of an aptamer-based direct-detection method for target molecules requiring molecular recognition using the thrombin and thrombin binding aptamer (TBA)³⁴ pair as a model system. TBA is a 15-mer ( $5^{\prime }$ -GGTTGGTGTGGTTGG- $3^{\prime }$ ) oligonucleotide that binds to the fibrinogen-recognition exosite of thrombin³⁹ after forming a chair-type G-quadruplex structure^{40, 41, 42, 43, 44} (Fig. 1 ). The compact size of the folded aptamer $(\sim 1.1\mathrm{nm})$ ⁴⁵ makes it an ideal receptor for this assay, as it brings the target molecule in close proximity to the metal surface, i.e., the analyte is well within the required distance $(<10\mathrm{nm})$ for local SERS field enhancement.

Fig. 1

(a) Antiparallel quadruplex or chair conformation of the TBA. (b) Guanine quartet of the TBA formed via the Hoogsten-type guanine-guanine interaction. In an antiparallel quadruplex, the guanosine residues have an alternating syn and anti conformation.

The technique (Fig. 2 ) utilizes immobilized nanoparticles that are functionalized with an aptamer. Binding of thrombin is detected by the presence of additional Raman bands following exposure of a TBA-functionalized SERS substrate to the test sample. The method described here may reduce the number of steps required in a sandwiched-based assay, and has the potential to discriminate between specific and nonspecific binding events.

Fig. 2

Scheme for direct detection of thrombin. (a) Nanoparticles are immobilized on a glass surface previously treated with (3-aminopropyl) trimethoxysilane. (b) Nanoparticles are functionalized with the thiolated thrombin binding aptamer. (c) The exposed metal surface is blocked by 6-mercaptohexanol. (d) The target molecule (thrombin in this case) is introduced. SERS spectrum is obtained at each step.

2.

Experimental Section

The general protocol for the assay is presented in Fig. 2. First, the silver nanoparticles are immobilized on a glass surface previously treated with 3-aminopropyl trimethoxysilane. The immobilized nanoparticles are functionalized with a thiolated thrombin binding aptamer and the surface is blocked with 6-mercaptohexanol to prevent nonspecific binding. A solution containing thrombin is then introduced. SERS spectra are obtained for each step to monitor the relevant binding events. Binding of thrombin is detected by the presence of additional Raman bands following exposure of TBA-functionalized SERS substrate to the test sample.

3.

Results and Discussion

The use of aptamers as molecular receptors has been attracting a lot of attention for a number of reasons. First, they are less susceptible to denaturation and degradation than the corresponding antibodies. Second, they can be synthesized with high purity and reproducibility and are also easily engineered. Third, they are much smaller than the corresponding antibodies, which is significant in applications where a smaller size of the capture ligand is desired. This is represented in the present case, where the target molecule needs to be in close proximity to the metallic nanoparticle for significant SERS enhancement.

The methodology reported here offers several potential benefits: 1. it does not require biotin-streptavidin binding, as the aptamer can be attached covalently to the SERS substrate via thiolchemistry; 2. the SERS substrate functionalized with aptamer is robust, owing to the stability of the aptamer; and 3. it is also compact and could be further miniaturized to make it suitable for field applications using portable Raman spectrometers.

We have previously reported a Raman and SERS characterization study of both the thiolated and unthiolated analogs of the thrombin binding aptamer (TBA).⁴⁷ We have shown that the formation of the quadruplex structure by TBA, indicated by the diagnostic $\sim 1480\text{-}{\mathrm{cm}}^{-1}$ band ( $\mathrm{C}8\mathrm{N}7–\mathrm{H}$ deformation, Fig. 1)^{48, 49, 50} under favorable conditions, i.e., presence of K+ and incubation at $\sim 4°\mathrm{C}$ , can be conveniently monitored by Raman spectroscopy. We also showed that the Raman and SERS spectra of the aptamer are dramatically different. This observation is not uncommon in SERS studies, and can be attributed to different enhancement factors of vibrational modes due to the fact that chemical bonds that are closer to the metallic nanoparticle surface receive stronger enhancement.

For this study, we have used a thiolated analog of the aptamer to facilitate binding of the aptamer to the nanoparticle. To show that thiolation does not prevent formation of the quadruplex structure, we previously measured the Raman spectra of both the thiolated and unthiolated analogs under the same conditions.⁴⁷ Raman spectra of the two analogs both exhibit the characteristic $1480\text{-}{\mathrm{cm}}^{-1}$ band of the quadruplex structure, indicating that thiolation does not inhibit the formation of the quadruplex structure by the aptamer, which is responsible for its binding to thrombin. To ensure that the aptamer is in its quadruplex form before immobilization, it was prepared in PBS containing ${\mathrm{K}}^{+}$ ions.

Figures 3 and 3 present the individual SERS spectra of the thiolated aptamer (TBA) and the blocking agent mercaptohexanol (MH), respectively. It can be observed that the characteristic $1480\text{-}{\mathrm{cm}}^{-1}$ band of the quadruplex aptamer structure is missing in the SERS spectrum. This may suggest that the aptamer is in its unfolded form, but the absence of the band around $1496{\mathrm{cm}}^{-1}$ , which corresponds to the $\mathrm{C}8\mathrm{N}7$ vibration of guanine when it is not hydrogen bonded,^{50, 51} indicates otherwise. It is very likely that the guanine tetrads do not interact with the metal, since they are not exposed to it, i.e., their location in the quadruplex configuration does not allow for it. Another very noticeable feature of the SERS spectrum of TBA is the presence of a strong band around $1398{\mathrm{cm}}^{-1}$ . This can be attributed mostly to the $\mathrm{C}{\mathrm{H}}_2$ deformation,⁵² which is shifted from the $1425\text{-}{\mathrm{cm}}^{-1}$ band of the spontaneous Raman spectrum,⁴⁷ with some contribution from the ring-stretching mode of the guanine $(\sim 1392{\mathrm{cm}}^{-1})$ .⁵⁰ The intensity of the peak suggests there is proximity of this moiety to the metal surface.

Fig. 3

SERS spectra of (a) immobilized TBA, (b) blocking agent (MH), and (c) immobilized TBA together with MH. The spectra are normalized with respect to the most intense peak for each trace.

Figure 3 shows the SERS spectrum of TBA together with MH. It can be seen that the major SERS peaks corresponding to MH (1080 and $1435{\mathrm{cm}}^{-1}$ , C–C stretch and C–H deformation, respectively) added to the substrate already functionalized with the aptamer are detected together with those peaks corresponding to the aptamer (1005, 1245, and $1398{\mathrm{cm}}^{-1}$ ). The addition of the target protein (thrombin) resulted in additional bands in the spectra [Fig. 4 ], namely $1635{\mathrm{cm}}^{-1}$ (amide 1, $\mathrm{C}\mathrm{O}$ stretch coupled with C–N–H bend),^{53, 54} $1540{\mathrm{cm}}^{-1}$ (amide 2, N–H bend coupled with C–N stretch),^{55, 56} and $1140{\mathrm{cm}}^{-1}$ (C–N stretch),⁵⁴ which are protein characteristic vibrational peaks (Table 1 ). In addition, there is a three-fold increase in the intensity of the C–H peak upon addition of thrombin. These observations strongly suggest binding of the target molecule. It may be noted that the amide 2 band is not normally Raman active. However, Raman enhancement near metallic nanoparticles can result in the appearance of vibrational modes that are normally forbidden by selection rules, as well as in the disappearance of certain vibrational bands. This is due to several factors, which include the proximity of the molecular bond to the metal surface and different orientation of the molecule with respect to the metal surface following surface adsorption or binding to capture agent. This may explain the appearance of this band at $1540{\mathrm{cm}}^{-1}$ . Similar observations have been reported in the literature. For example, Hu ⁵⁶ reported the appearance of an amide 2 band (centered at $1526{\mathrm{cm}}^{-1}$ ) in the SERS spectrum of lysozyme, Kim ⁵⁷ observed the presence of this band (at $1520{\mathrm{cm}}^{-1}$ ) in the SERS spectrum of Gly-L-Phe dipeptide, and Podstawka ^{58, 59} observed this band from $1504\text{to}1554{\mathrm{cm}}^{-1}$ in the SERS spectra of a series of homo- and heterodipeptides.

Fig. 4

(a) SERS of immobilized TBA together with MH (solid line) and TBA with MH and the target protein, thrombin (dashed line). The spectra normalized with respect to the $1398\text{-}{\mathrm{cm}}^{-1}$ TBA peak. (b) SERS spectra of immobilized TBA with MH (solid line) and immobilized TBA with MH and the nonbinding BSA (dashed line). The spectra are normalized with respect to the most intense peak $\left(2920{\mathrm{cm}}^{-1}\right)$ for both.

Table 1

SERS frequencies of TBA, TBA+MH , and TBA+MH+thrombin and their corresponding vibrational modes. The observed additional peaks on TBA+MH and TBA+MH+thrombin spectra are listed in bold for clarity.

These spectral changes that were observed with the addition of thrombin to the aptamer functionalized substrate were not detected when the same experiment was conducted using a nonbinding protein (BSA) [Fig. 4]. The spectra of TBA with MH and of TBA with MH and BSA are almost superimposable, suggesting that no binding has occurred. However, it can also be seen that there is a noticeable change in the spectrum of TBA with MH, specifically the relative intensities of their major peaks. One possible explanation for this observation is the variation in the amounts of aptamer and mercaptohexanol that were bound to the nanoparticle. The intensity ratio of $1398\text{-}{\mathrm{cm}}^{-1}∕1080\text{-}{\mathrm{cm}}^{-1}$ peak is noticeably higher in this set of spectra, suggesting an increased amount of aptamer adsorbed on the nanoparticle compared to the previous experiment (using thrombin). It is likely that this increase in the amount of aptamer molecules bound to the nanoparticle surface resulted in a higher relative intensity of the C–H peak. It is also possible that the higher concentration of bound aptamer led to a change in its orientation with respect to the nanoparticle surface, i.e., from flat or slightly bent to perpendicular, which in turn results in the observed higher relative intensity of the C–H peak.⁶⁰ It has been shown both experimentally and theoretically that the C–H stretching mode is sensitive to molecule orientation on the nanoparticle surface.^{61, 62} In addition to the already mentioned changes in the spectrum of BSA-NP complex, we noted an additional band around $1518{\mathrm{cm}}^{-1}$ . One possible explanation for the appearance of this peak is nonspecific binding of BSA or other components (e.g., buffer components) of the test sample to the nanoparticle, resulting from insufficient coverage of the metal surface by both TBA and MH. This observation may indicate that nonspecific binding can be detected and discriminated against based on the different spectral response. While the sensitivity is not yet comparable with commonly used label-based and sandwiched-type detection methods such as ELISA, it is comparable with other direct detection techniques available in the literature, such as those reported by Jung,⁶³ a surface acoustic wave (SAW)-based detection of thrombin $\left(1\mu \mathrm{M}\right)$ using an aptamer-functionalized sensor chip; and by Neumann,¹⁵ a SERS-based direct detection of target molecules (PDGF and cocaine), wherein the randomness of the SERS spectra of the specific aptamer following exposure to the corresponding target molecule ( $1\mu \mathrm{M}$ PDGF and $20\mu \mathrm{M}$ cocaine) is monitored. We believe that further improvements (e.g., preconcentration of the sample on the SERS substrate) to the current technique will enhance its sensitivity.

4.

Conclusions

We demonstrate a simple and selective SERS-based method for direct detection of biomolecules based on molecular recognition, which require fewer steps than traditional sandwiched-based assays. Binding of the target molecule to the molecular receptor (aptamer) is manifested by the appearance of the molecule’s characteristic vibrational bands (C–N stretching, amide 1 and amide 2 bands), in addition to the vibrational peaks corresponding to the receptor and blocking agent. It is important to note that interfering species that are able to bind may produce a noticeably different spectral response, thus further improving the selectivity.

Since aptamers are available not only for biomolecules but also for small molecules such as ricin, trinitrotoluene, and polycyclic aromatic hydrocarbons, this technique can be useful for the detection not only of medical biomarkers but also of environmental pollutants, as well as biological and chemical threat agents. In addition, the SERS substrates can be functionalized with different aptamers, allowing for multiplexed screening and detection. The simplicity of the developed label-free detection technique could be further utilized for on-chip and real-time detection by implementing it on a nano- and microscale platform.