Bioluminescent Protein—Inhibitor Pair in the Design of a Molecular Aptamer Beacon Biosensing System

Abstract

Although bioluminescent molecular beacons designed around resonance quenchers have shown higher signal-to-noise ratios and increased sensitivity compared with fluorescent beacon systems, bioluminescence quenching is still comparatively inefficient. A more elegant solution to inefficient quenching can be realized by designing a competitive inhibitor that is structurally very similar to the native substrate, resulting in essentially complete substrate exclusion. In this work, we designed a conjugated anti-interferon-γ (IFN-γ) molecular aptamer beacon (MAB) attached to a bioluminescent protein, Gaussia luciferase (GLuc), and an inhibitor molecule with a similar structure to the native substrate coelenterazine. To prove that a MAB can be more sensitive and have a better signal-to-noise ratio, a bioluminescence-based assay was developed against IFN-γ and provided an optimized, physiologically relevant detection limit of 1.0 nM. We believe that this inhibitor approach may provide a simple alternative strategy to standard resonance quenching in the development of high-performance molecular beacon-based biosensing systems.

Nucleic acid biosensors can provide highly selective target recognition in the diagnosis of infectious diseases,¹ identification of genetic alterations,² and, in the case of aptamers, the detection of clinically relevant non-nucleic acid molecules.³ The popularity of nucleic acid biosensors have recently increased due to their ability to merge highly sensitive and selective detection, thermodynamic stability, and a broad range of functionalization with modern point-of-care demands such as simplistic assay interfaces and rapid generation of results.⁴ Of particular importance to assay design, nucleic acid probes are easily synthesized, highly selective, and can detect practically any target of interest, from DNA and RNA to proteins and small molecules.⁵

Molecular beacon (MB) technology, first developed by Tyagi and Kramer in 1996, is a variant of the traditional hybridization-based fluorescence quenching assay platform in which the separate and independent donor/acceptor resonance energy transfer probes were replaced with a single-stranded oligonucleotide that self-hybridizes in a stem-loop configuration.⁶ Upon hybridization of a complementary target to the single-stranded loop region, the double-stranded stem is forced apart. By conjugating a fluorophore/quencher pair to the stem termini, target-induced opening of the stem removes the quencher from close proximity of the fluorophore, leading to a dose-dependent increase in fluorescence. This type of fluorescent MB has been used in many applications besides nucleic acid detection, including real-time hybridization in living cells,^7,8 DNA–protein interactions,⁹ monitoring enzymatic cleavage,¹⁰ and real time polymerase chain reactions (qPCR).^11,12 However, they can suffer from severe sensitivity issues mainly due to background fluorescence from the excitation source.⁵ In order to address these shortcomings, our group was the first to report a bioluminescent stem-loop probe (BSLP).^13,14 For this, a bioluminescent protein (Renilla luciferase, Rluc8) replaced the traditional fluorophore and demonstrated a significant improvement in sensitivity. Modulation of the bioluminescent signal was essentially identical to a standard MB and was applied to the detection of miR-21, a commonly dysregulated biomarker. Our laboratory also developed a BSLP using Gaussia luciferase (Gluc) rather than Rluc8 in order to improve sensitivity, as Gluc is smaller and has stronger bioluminescence activity. Although a drastic improvement in the signal amplitude for both systems was demonstrated when compared to fluorescence-based stem loop probes, similar drawbacks to a standard MB such as inefficient quenching of the bioluminescent signal were observed, resulting in a signal-to-noise ratio that could still be improved.

We hypothesized that another, more elegant, solution to inefficient quenching could be realized by capitalizing on the unique enzyme/substrate relationship of a specific bioluminescent protein. By designing a competitive inhibitor that is structurally very similar to the native substrate, essentially complete substrate exclusion can be realized. We hypothesized that a significant improvement in MB sensitivity can be achieved by eliminating unintended triggering of the bioluminescent reaction altogether. This method relies on preferential occupation of the protein active site by the inhibitor while the MB is in the closed conformation, thereby preventing enzymatic oxidation of the luminescent substrate. Target binding provides sufficient strain to the MB structure such that the substrate analogue is forcibly separated from the active site, allowing the resumption of normal substrate oxidation. Several detection systems have already employed inhibitors capable of interacting with the luciferase active site with varying degrees of affinity.¹⁵ Poutiainen et al.¹⁶ developed a peptide-based sensor conjugated with a luciferase inhibitor, and Shimada et al.¹⁷ designed a weakly inhibiting aptamer system for thrombin detection. In another study by Schena et al.,¹⁸ a construct was developed between a luciferase and carbonic anhydrase where a synthetic ligand with two mutually exclusive binding sites modulates the activity of the reporter externally as opposed to inducing a conformational change. To the best of our knowledge, however, there are no inhibitor-based MBs that combine a bioluminescent protein with a substrate analogue. Importantly, it has been shown elsewhere that the effective concentration of an inhibitor dramatically increases upon local tethering to a target enzyme, thereby achieving inhibition at lower concentrations than a free inhibitor.¹⁹

To extend this novel quenching method to the detection of non-nucleic acid targets, we developed an aptamer-based bioluminescent MB. Aptamers are single-stranded oligonucleotides with binding properties similar to antibody—antigen interactions.^20,21 They are identified by a method called Systematic Evolution of Ligands by Exponential enrichment (SELEX) that amplifies only those nucleic acid sequences that bind specific targets such as proteins, small molecules, and even whole cells.^22,23 The advantages of using aptamers over antibodies include simple synthesis, easy labeling with fluorophore dyes or enzymes, and greater stability in nonphysiological conditions.²⁴ The creation of a molecular aptamer beacon (MAB) combines the advantages of a MB such as binding specificity and switchable reporter signaling with a system that can recognize diverse targets without a requirement for hybridization. In addition, this construct could provide a selective sensing element to a large variety of platform designs, enabling the merger of discrete components or an overall simplification of sensor design.^25–27 This manuscript describes the combination of an inhibition-based MB with an aptamer-based loop domain in order to detect the immune-stimulating cytokine, interferon-γ (IFN-γ) (Figure 1). MABs have been developed in the recent past for real-time detection,^28,29 imaging,³⁰ and other applications.³¹ However, this is the first demonstration, to our knowledge, of a bioluminescent aptamer-based MB and is certainly the first use of an inhibitor rather than a quencher in MAB design. This strategy provides an innovative and highly sensitive addition to the mix-and-measure toolkit of molecular beacon technology. In this work, the inhibitor molecule coelenteramine was chosen for the bioluminescent protein Gaussia luciferase (GLuc) due to the structural similarity to its native substrate, coelenterazine. By designing a competitive inhibitor that is structurally very similar to the native substrate, essentially complete substrate exclusion can be realized. We hypothesized that a significant improvement in MB sensitivity can be achieved by eliminating unintended triggering of the bioluminescent reaction altogether.

Figure 1.

General principle of the bMAB assay. Binding of the target results in the conformational change in the MAB, allowing separation of the tethered inhibitor from Gluc and subsequent binding of the substrate. While the IFN-γ crystal structure has been resolved (PDB 1HIG), an example protein structure has been included to represent Gluc, as the crystal structure is unknown.

RESULTS AND DISCUSSION

Design of a Molecular Aptamer Beacon (MAB).

The anti-IFN-γ sequence that formed both the 5′ stem region of the MAB as well as the majority of the loop (5′-GGGGTTGGTTGTGTTGGGTGTTGTGTC) has been well-characterized in the literature,³² while the 3′ stem was added as an antisense sequence to the first seven nucleotides of the aptamer sequence. The oligonucleotide sequence was designed with a 5′-carboxylic acid for coelenteramine-inhibitor conjugation and a 3′-azide for bioconjugation of GLuc. The coelenteramine-inhibitor was synthesized in a two-step reaction, beginning with the addition of a polyethylene glycol (PEG) spacer on the coelenteramine molecule and followed by deprotection of the primary amine moiety (Figure 2A). The PEG spacer allowed flexibility of the inhibitor to enhance binding interactions with GLuc. For comparison of inhibition efficiency, we synthesized PEG spacers with either two or four ethylene oxide monomers. Both products were characterized using NMR spectroscopy (Figures S6–S10).

Figure 2.

(A) Two-step reaction of the coelenteramine-inhibitor synthesis. (B) Bioconjugation scheme of the bioluminescent MAB. (C) SDS-PAGE gel stained with silver staining: lane 1, protein ladder; lane 2, purified GLuc; and lane 3, conjugated MAB with GLuc and coelenteramine-inhibitor.

The general bioconjugation scheme for the bioluminescent MAB (Figure 2B) involved modifying surface-exposed lysines on Gluc with dibenzocyclooctyne-N-hydroxysuccinimidyl ester (NHS-DBCO; Lumiprobe, Hunt Valley, MD) in order to introduce a reactive cyclo-alkyne molecule. The formation of the product, GLuc-DBCO was confirmed by UV–visible spectroscopy using a DBCO-reactive azide-fluorophore 545 (Sigma, St. Louis, MO). The absorption spectrum between 350 and 650 nm confirmed two characteristic peaks around 520 and 550 nm (Figure S6B). Concurrently, the MAB was conjugated with coelenteramine-inhibitor at the 5′-carboxylic acid end via 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide (EDC) activation and hydroxysulfosuccinimide (sulfo-NHS) stabilization of the amine-reactive intermediate. Because the coelenteramine-inhibitor is electroactive, modification of the MAB was monitored using differential pulse voltammetry based on a positive shift in voltage potential following conjugation (Figure S7). Finally, the 3′-azide end of the stem was conjugated to the GLuc-DBCO through copper-free click chemistry. The decision to use copper-free click chemistry was based on a literature demonstration of potential detrimental effects of copper ions on biomolecules³³ and proteins.³⁴ This final bioconjugation reaction was performed on a streptavidin–biotin column using a biotinylated complementary oligonucleotide to the aptamer sequence for the attachment of the probe to the resin-bound streptavidin. Besides enabling simple removal of unconjugated Gluc, this method forced the MAB into the open configuration to avoid any steric hindrance between the protein and the stem of the MAB. The formation and purity of the conjugated bMAB was confirmed using SDS-PAGE analysis (Figure 2C). To minimize product loss during characterization, a small aliquot was analyzed and imaged using silver staining (Thermofisher, Weston, FL) according to the manufacturer’s instructions. As shown in Figure 2C, the characteristic band of GLuc appears in lane 2 at approximately 22 kDa, while the conjugated MAB showed a major band around 37 kDa that corresponded to a monoconjugated Gluc. Some lack of visible definition for the conjugate bands likely resulted from variable DBCO conjugation stoichiometry. Smearing at higher molecular weights was partially the result of purification artifacts visible after long staining times but also served to demonstrate the presence of polyconjugates that were anticipated due to the likelihood of multiple surface-exposed lysines available for conjugation. For convenience, the products with different PEG linkers will be referred as aptamer(PEG2)-GLuc and aptamer-(PEG4)-GLuc in this manuscript (Figure S1).

Optimization of Assay Conditions.

In order to demonstrate that an inhibitor-based bMAB design can achieve higher sensitivity and a better signal-to-noise ratio, we developed a bioluminescent assay for the detection of IFN-γ. Initially, the temperature and the incubation time of the assay were evaluated under conditions likely to be encountered for physiological sampling. We optimized the stem portion of our molecular aptamer beacon for hybridization at or below room temperature (20 °C) in order to provide optimal in vitro sensing capabilities. As the melting temperature of the stem is the most crucial design aspect relative to the signal-to-noise ratio (S/N), we specifically tailored the stem for the best S/N possible within this design for operation within a narrow temperature window. This was a necessary feature, as robust stem hybridization requires higher assay temperatures, limiting the choice of instrumentation, reducing target affinity, and exceeding the optimal kinetic parameters of Gluc. For temperature optimization, a comparison of assay performance at 4 °C and room temperature (~20 °C) indicated that the sensitivity of the system was greater at 20 °C (Figure S8). As expected, lower than optimal temperatures result in incomplete stem denaturation and poor overall detection limits. Out of curiosity, we did attempt a dose–response curve at 37 °C but found complete signal saturation due to the lack of stem-loop hybridization at any IFN-γ concentration. Moreover, incubation times were investigated for a single IFN-γ concentration at the extreme lower end of the detection range. For this optimization study, the conjugated MAB solution was mixed with 1.0 nM of IFN-γ, and the mixture was incubated for 0.5, 1.0, and 2.0 h. The signal-to-noise ratio (S/N) was calculated as the ratio of bioluminescent signal at a specific concentration of the analyte over the signal of the background without any analyte present. We chose 1.0 h incubation time because it resulted in better S/N when compared to the 30 min or 2 h time points. In comparing the dose–response curves for aptamer(PEG2)-GLuc and aptamer(PEG4)-Gluc, we concluded that the aptamer(PEG4)-GLuc has better S/N (Figure S9) and was, therefore, chosen as the representative bMAB for further studies.

Comparison of MAB with Inhibitor and Quencher.

To demonstrate that the inhibition-based bMAB gives a better S/N ratio when compared to the quenching-based method, we replaced the inhibitor molecule with the standard quencher molecule, DABCYL, while using the same oligonucleotide sequence. Bioluminescence responses were recorded for equimolar MAB and IFN-γ concentrations within the same experimental run, and Figure 3 demonstrates that the S/N ratio for all IFN-γ concentrations was higher when using an inhibitor rather than a quencher. Additionally, when comparing the calibration curves, the system response for the inhibitor-MAB is 3.6-fold better than the system response for the quencher-MAB.

Figure 3.

Comparison of (A) the dose response curves and (B) S/N ratios between inhibition and quenching. Data points are an average of three measurements ±1 standard error of the mean. In some cases, the error bars are obscured by the symbols of the points.

Calibration Curve.

After optimization, concentrations of IFN-γ ranging from 5 × 10⁻¹¹ to 1 × 10⁻⁶ M were tested against 15 nM of bMAB, and the measured data was plotted using GraphPad Prism software to calculate the quantification range of the bioluminescent assay. From 1 μg/mL substrate injections, a calibration curve was generated from the dose–response curve (Figure 4), providing a calculated detection limit of 1.0 × 10⁻⁹ M IFN-γ for the aptamer(PEG4)-GLuc bMAB. This limit of detection was interpolated using the formula LOD = S_B + 3 × SD_B where S_B and SD_B are the mean and the standard deviation of the blank signals, respectively. This LOD was significantly biased by the IFN-γ binding constant of the aptamer sequence, which was reported to be 3.0 nM by Tuleuova et al.³⁵

Figure 4.

Dose–response curve for IFN-γ. The data points are an average of three measurements ±1 standard error of the mean. Some error bars are obscured by the symbols of the points.

Specificity of the bMAB.

The specificity of the bMAB was studied against two cytokines, interleukin-6 (IL-6) and tumor necrosis factor alpha (TNF-α), that are similarly produced by the immune system as biological responses against disease.³⁶ As shown in Figure 5, essentially no bioluminescence response against either protein was evident, indicating that the aptamer(PEG4)-Gluc bMAB can be considered unlikely to provide false-positive signals within these assay conditions.

Figure 5.

Specificity test against two different cytokines. The data points are an average of three measurements ±1 standard error of the mean. Some error bars are obscured by the symbols of the points.

CONCLUSIONS

By conjugating a bMAB with a noncleavable substrate analogue that inhibits GLuc bioluminescence prior to target binding, we have generated, to the best of our knowledge, the first example of a dose-responsive, inhibitor-based bioluminescent MAB. Inhibition was provided by coelenteramine, a downstream degradation product of the standard bio-luminescent substrate of Gluc, coelenterazine. Conjugation of Gluc and its inhibitor to opposite ends of a MAB allowed us to develop a solution-phase assay against a pro-inflammatory immune molecule, IFN-γ. This bioluminescence assay demonstrated a better analytical response and signal-to-noise ratio than a quencher-based bMAB using DABCYL. Moreover, the bMAB showed excellent specificity against other cytokines. We believe that the enhanced performance of an inhibitor bMAB represents a significant improvement in molecular beacon design and could be directly applied to a variety of traditional and aptamer beacon assays.