Utilization of chromogenic enzyme substrates for signal amplification in multiplexed detection of biomolecules using surface mass spectrometry

Abstract

MicroRNAs (miRNAs) are important post-transcriptional gene regulators and can serve as potential biomarkers for many diseases. Most of the current miRNA detection techniques require purification from biological samples, amplification, labeling, or tagging, which makes quantitative analysis of clinically relevant samples challenging. Here we present a new strategy for the detection of miRNAs with uniformity over a large area based on signal amplification using enzymatic reactions and measurements using time-of-flight secondary ion mass spectrometry (ToF-SIMS), a sensitive surface analysis tool. This technique has high sequence specificity through hybridization with a hairpin DNA probe and allows the identification of single-base mismatches that are difficult to distinguish by conventional mass spectrometry. We successfully detected target miRNAs in biological samples without purification, amplification, or labeling of target molecules. In addition, by adopting a well-known chromogenic enzymatic reaction from the field of biotechnology, we extended the use of enzyme-amplified signal enhancement ToF (EASE-ToF) to protein detection. Our strategy has advantages with respect to scope, quantification, and throughput over the currently available methods, and is amenable to multiplexing based on the outstanding molecular specificity of mass spectrometry (MS). Therefore, our technique not only has the potential for use in clinical diagnosis, but also provides evidence that MS can serve as a useful readout for biosensing to perform multiplexed analysis extending beyond the limitations of existing technology.

Graphical Abstract

We describe a novel, high-throughput method to detect macromolecular interactions. Using time-of-flight secondary ion MS, we detected insoluble mass signal enhancer molecules generated by the interaction between routinely used reporter enzymes and their substrates in presence of target analytes. Our method bypasses the technical limitations of common multiplex assays and has potential applications in studying a broad range of biological phenomena.

Introduction

MicroRNAs (miRNAs) are involved in a variety of biological and pathological events in living organisms, which makes the detection of these biomarkers extremely important (Ambros, 2004; Bartel, 2004; Kanwal et al. 2017; Liu et al. 2012; Ruan et al. 2009). However, owing to their low abundance, fragile structure, and sequence similarity, miRNAs are notoriously difficult to detect (Huang et al. 2015; Hunt et al. 2015; Labib et al. 2013). For this reason, miRNA detection platforms should be able to perform multiplex analysis with high sensitivity and sequence specificity, without labeling the target (Joshi et al. 2017; Ryoo et al. 2013; Xiao et al. 2009). Numerous techniques have been developed in an attempt to meet these requirements, including polymerase chain reaction (PCR), next-generation sequencing (NGS), fluorescence resonance energy transfer (FRET), electrochemical-based, surface-enhanced Raman scattering (SERS), and localized surface plasmon resonance (SPR)-based methods (Na et al. 2018; Tian et al. 2015; Xiao et al. 2009; Xue et al. 2019). Despite these efforts, however, a number of constraints such as cost, labor intensive processes, reproducibility, quantification, sensitivity, and multiplexing, have prevented these platforms from being clinically feasible. Moreover, as most technologies are based on optical readout, it is difficult to precisely detect target molecules with molecular specificity that takes into consideration the elemental compositions and molecular structures. Among the various analytical techniques, mass spectrometry (MS) is a particularly useful tool, as it provides information regarding the molecular weight (Mrksich, 2008; Parker and Borchers 2014). As it can distinguish a difference of a few daltons in molecular weight with unique isotope patterns, which is the inherent characteristic of a substance, MS has great potential to provide most decisive information and be used in multiplex assays without requiring analyte labeling (Kim et al. 2012). MS techniques are numerous; of these, time-of-flight secondary ion MS (ToF-SIMS) is a notably powerful surface analysis tool that enables highly reproducible and sensitive measurement and imaging of molecules present on surfaces without requiring matrix treatment, a process that often causes difficulties during sample preparation and data analysis (Carlred et al. 2014; Kim et al. 2007; Nuñez et al. 2018; Van Nuffel et al. 2018; Yokoyama et al. 2016).

Biomolecular detection is widely acknowledged to be a crucial part of the diagnostic process; however, understanding the interactions between biomolecules is also of considerable biological significance (Cabezas et al. 2017; Lee et al. 2008; O’Kane and Mrksich, 2017). In this regard, although MS techniques have contributed greatly to the identification of biomolecules, they are unsuitable for the analysis of non-covalent protein-protein and DNA-RNA interactions or for the detection of low-abundant biomolecules from most biological samples (Kullolli et al. 2014). In addition, identification of biomolecules using MS often requires separation of the analyte from the biological mixtures as well as its digestion (Ma et al. 2014). Indeed, attempts have been made toward achieving mass tag-based signal amplification to circumvent the limitations of MS in biomolecular detection, but these attempts have been primarily based on extensive modifications such as conjugation of tag molecules or nanoparticles to the antibodies, DNAs, and RNAs, which can affect the affinity of interaction with the target (Angelo et al. 2014; Lorey et al. 2015; Tayyari et al. 2013). Therefore,, there is a compelling need for the development of novel molecular signal enhancers (MSEs) with increased sensitivity to target signals and high target specificity without the need for any modification (Jiang et al. 2017). Especially, because the current miRNA detection techniques based on MS require additional steps for gene amplification by PCR, involving reverse transcription (RT), labeling (including isotopic labeling), or sequencing (Kim et al. 2016; Lathwal and Sikes, 2016; Wambua et al. 2014), it is highly desirable to develop an efficient method for sensitive and specific miRNA detection in biological samples without these steps.

Here, we describe a versatile and powerful strategy that involves the use of chromogenic enzymes to form insoluble products that act as MSEs in the surface MS techniques used for miRNA detection. As one equivalent of a bound enzyme (resulting from one recognition) is converted into multiple equivalents of MSEs from an excessive amount of soluble enzyme substrates, signal amplification is achieved using various conventional chromogenic enzymes for the ultrasensitive detection of miRNAs in a high-throughput format with excellent reproducibility. We achieved highly specific miRNA detection, enabling discrimination of single-base mismatches, on a gold chip-based assay platform. Furthermore, our enzyme-amplified signal enhancement ToF (termed EASE-ToF) strategy enabled a multiplex assay and the sensitive detection of miRNA in biological samples and allowed the application of matrix-assisted laser desorption/ionization (MALDI)-ToF and laser desorption/ionization (LDI)-ToF analyses. In addition, our strategy could be easily extended to protein detection by applying the sandwich type sensing platform. We believe this to be the first method to use surface MS for quantitative and sensitive miRNA and protein detections in biological samples without target amplification and labeling.

Materials and Methods

Materials.

DNA probes were purchased from Genotech (Daejeon, Korea). MicroRNAs were purchased from Bioneer (Daejeon, Korea). Phosphate-buffered saline (PBS), Dulbecco’s modified Eagle’s medium (DMEM), RPMI 1640 medium, and fetal bovine serum (FBS) were purchased from WelGene (Gyeongsan, Korea). Streptavidin-conjugated enzymes (HRP, AP, and β-gal) and their substrates (DAB, AEC, 4-CN, BCIP/NBT, and S-gal), as well as bovine serum albumin (BSA), 3-mercaptopropanol, 1-ethyl-3–3(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC), N-hydroxysulfosuccinimide (sulfo-NHS), and Grace Bio-Labs CultureWell^™ removable chambered coverglasses were purchased from Sigma-Aldrich (St. Louis, MO, USA). The monoclonal anti-interferon gamma antibody and biotin conjugated monoclonal antibody were purchased from Abcam (Cambridge, UK). The TRIzol reagent was purchased from Invitrogen (Carlsbad, CA, USA). All reagents were used as received.

Formation of self-assembled monolayers.

Gold wafers were prepared by vacuum deposition of a 15-nm-thick Cr film, followed by the deposition of a 100-nm-thick gold layer, onto a silicon wafer. The gold wafers were cleaned by treatment with a super-piranha solution (20 mL of H₂O₂, 2 mL of HNO₃, and 12 mL of H₂SO₄). To prepare a multi-well gold chip, 16-well CultureWell^™ coverglasses were introduced onto the gold wafers. Before the formation of a self-assembled monolayer (SAM), DNA probes were annealed at 95 °C for 5 min and gradually cooled down to 4 °C for 1 h in 10 μM concentration in PBS. A hairpin probe solution in PBS was added to the wells (100 μL per well) on gold chips with 10 μM of 3-mercapto-1-propanol. After overnight incubation (In most cases, incubation with thiol-modified molecules for 6 h is sufficient for SAM formation.), the chips were washed with PBS five times.

Treatment with miRNA and enzyme reactions on gold chips.

The gold chips modified with probes were treated with predetermined concentrations of miRNAs in RNase-free PBS (or in total RNA dissolved in PBS) at room temperature for overnight. Then, the mixture was incubated at 55 °C for 1 h, followed by washing with PBS three times at 55 °C for 10 min. After washing, the gold chips in PBS were cooled to 4 °C. The biotinylated signaling probe, whose sequence is shown in Table S1, was dissolved in PBS at a concentration of 200 nM and applied to each well on the gold chip. After incubation for 3 h, the gold chips were washed with PBS five times and incubated with 1% BSA in PBS for 1 h. A streptavidin-AP conjugate solution with 1% BSA in PBS was added to the sensor at a concentration of 5 μg/mL for 1 h. After repeated washing with PBS, the gold chips were incubated with a premixed BCIP/NBT solution for 10 min following the manufacturer’s protocol. Enzymatic reactions with other kinds of substrates were also performed according to the manufacturer’s instructions.

ToF-SIMS measurements.

All measurements were acquired using a ToF-SIMS V instrument (IONTOF GmbH, Germany). Spectra and images were obtained using a Bi₃⁺ as an analysis ion beam with approximately 0.05 pA at a cycle time 150 μs. For the spectra, the analysis area was 250 μm × 250 μm and primary ion dose density (PIDD) was 1.0 × 10¹¹ ions/cm² to ensure static limits. For an image with a wide field of view of 54.2 mm × 16.3 mm, the stage-scan method was performed in 190 × 58 patches in a random mode. Patch size was approximately 300 μm × 300 μm and PIDD was 1.15 × 10⁹ ions/cm². The negative ion spectra were mass-calibrated using C₂H⁻, C₃H⁻, C₄H⁻, C₅H⁻, and C₇H⁻ peaks. The positive ion spectra were mass-calibrated using CH₃⁺, C₂H₃⁺, C₃H₅⁺, and C₅H₇⁺ peaks.

Results and Discussion

Figure 1 illustrates an EASE-ToF strategy based on precipitate formation by chromogenic substrates of enzymes for ultrasensitive miRNA detection using ToF-SIMS. In detail, hairpin DNA probes on gold chips were designed to be hybridized with target miRNAs, followed by hybridization with the same number of biotinylated signaling probes. Due to the presence of the biotinylated probes, the same number of streptavidin-enzyme conjugates were also bound selectively to the hairpin DNA probes on the gold surface. From an excess of soluble substrates the bound enzymes converted a sufficient number of soluble substrates into insoluble precipitates required for detection by ToF-SIMS. This protocol transduced the biological binding events into highly amplified mass signals through enzymatic reactions in a quantitative manner.

Figure 1.

Signal amplification strategy of MSE-MS for ultrasensitive miRNA detection using enzyme reactions and ToF-SIMS. The hairpin DNA probe (pink) on a gold chip is designed such that it can hybridize with the unlabeled target miRNA (green), followed by hybridization with a biotinylated signaling probe (red). Streptavidin-enzyme conjugates are bound to biotin and a soluble substrate is converted to an insoluble product (as a molecular signal enhancer, MSE) in an enzymatic reaction. The resulting organic molecules are analyzed using ToF-SIMS

To demonstrate signal amplification using MSEs produced by enzymes in a multiplexed manner, we used alkaline phosphatase (AP), horseradish peroxidase (HRP), and β-galactosidase (β-gal), which are chromogenic enzymes commonly used in various biological analyses, such as western blotting and immunohistochemistry (De Jong et al. 1985; Krieg and Halbhuber, 2010).

First, we prepared gold chips with immobilized hairpin DNA probes, using thiol-modified DNA with a sequence complementary to miR-let-7a. As the formation of mixed self-assembled monolayers (SAMs) with hydroxyl terminated alkyl chains is known to prevent non-specific adsorption of probe DNA, and have advantages with respect to proper orientation that enhance hybridization (Lee et al. 2006; Lee at al. 2007), we prepared mixed SAMs of 3-mercaptopropanol and DNA probes. Hybridization with miR-let-7a triggered the binding of a biotinylated signaling probe, followed by the binding of streptavidin-HRP, streptavidin-AP, or streptavidin-β-gal conjugate as chromogenic enzymes. Adding the soluble HRP, AP, or β-gal substrates, 3-amino-9-ethylcarbazole (AEC, Figure 2a), 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium (BCIP/NBT, Figure 2b), or 3,4-cyclohexenoesculetin-β-D-galactopyranoside (S-gal, Figure 2c), respectively, resulted in the formation of insoluble products on the surface of the gold chips in the presence, but not in the absence of miR-let-7a (Figure S1a). Using the ToF-SIMS technique, each enzyme reaction product was detected distinctively (Figure 2a, 2b, and 2c) depending on its chemical structure. ToF-SIMS spectra were obtained using a Bi₃⁺ as an analysis ion beam. In Figure 2b, a distinct peak was detected in the negative ion spectra at m/z 748.3, corresponding to intact NBT formazan, [M]⁻, which is produced by the hydrolysis of BCIP by AP and the subsequent reduction of NBT by the hydrolyzed product (Figure S2a and S2b) (De Jong et al. 1985). After normalization of peak intensities at m/z 748.3 (which were obtained in the presence of serially diluted miR-let-7a in PBS) to total ion counts, we observed an excellent linear relationship between the miR-let-7a concentrations and normalized peak intensities (Figure 2d). The limit of detection (LOD) was calculated as 8 aM (attomolar) using the following equation: LOD = 3.3 (SD/S), where SD is the standard deviation and S is the slope of the calibration curve in Figure 2d. In this analysis, the coefficient of variations (%CV) for intra- and inter-assay averages were 6.8 and 5.6, respectively. Using our signal amplification strategy based on an MSE produced by the enzyme, we achieved ultrasensitive detection of the miRNA with ToF-SIMS, to levels comparable to those achieved in previous reports that used a number of analytical methods such as SERS and single-molecule FRET (Johnson and Mutharasan, 2014). Utilization of the insoluble enzyme reaction products as mass signal transducers (Figure S1b, water-insoluble film was formed) might have contributed to the high sensitivity by the removal of salts and molecules on the surface that can interfere with MS measurement by washing with water in the last step, as well as stoichiometry between the analyte and the mass signal. Furthermore, as this approach is based on the recognition of non-covalent DNA-RNA interactions, it has the potential to be used in other biological assays such as protein-protein, protein-substrate or RNA-protein interactions using surface MS techniques (Chen et al. 2016).

Figure 2.

Detection of insoluble precipitate produced by the enzymatic reactions. ToF-SIMS spectra of enzymatic reaction products of a) horse-radish peroxidase (HRP), b) alkaline phosphatase (AP), and c) β-galactosidase (β-gal) in the presence of miR-let-7a. d) Structure of NBT-formazan (inset) and relationship between normalized intensity and the concentration of miR-let-7a. Data points represent the mean ± standard deviation (SD).

Next, we examined our strategy’s ability to discriminate single-base mismatches in miRNAs. Sequence specificity is crucial for miRNA-sensing platforms, as miRNAs are short and quite diverse with respect to type and function. Notably, the let7 miRNA family has numerous members with various functions (Hilly et al. 2016; Roush and Slack, 2008). As the sequences of these miRNAs are highly conserved, they were ideal candidates to demonstrate the specificity of our sensing platform. We chose miR-let-7a as a model and compared its detection efficiency with those of miR-let-7c and miR-let-7f, which contain single-base mismatches at different sites in their sequences (Figure 3b and Table S1). Sequence specificity was successfully achieved by thermal treatment at 55 °C to remove partially matched miRNA sequences, which hybridized in a less stable manner with the probe.

Figure 3.

Sequence-specific miR-let-7a detection (a–c), and miR-200a-3p detection in total RNA extracts from human primary gastric cancer cell lines (d and e). a) Spectra measured by ToF-SIMS in the absence (red) and presence of miR-let-7a (blue). b) Sequences of miRNAs. Mismatch positions in miRNA sequences are underlined and in bold. c) Intensity of the peak at m/z 748.3 in the presence of miR-let-7a, miR-let-7c, and miR-let-7f. d) Spectra measured by ToF-SIMS in the presence of 1 nM miR-200a-3p in buffer (blue) and in total RNA extract from a human gastric cancer cell line, SNU1 (red). e) Relationship between normalized intensities of the peak at m/z 748.3 measured by ToF-SIMS (bars and right y-axis) and miR-200a-3p expression analyzed by RT-PCR (dots and left y-axis) in two human gastric adenocarcinoma cell lines, SNU1 and MKN45. Data points represent the mean ± SD.

In Figure 3a and 3c, the intensities of the peaks at m/z 748.3 clearly show discrimination of single-base mismatches. An intense peak was observed in the presence of miR-let-7a, while peak intensities were negligible in the presence of miR-let-7c and miR-let-7f, demonstrating that our strategy enables discrimination of single-base mismatches without labeling, RT, gene amplification, or sequencing. Mismatches in miRNA sequences may include one or more of the four nucleotides or changes in their positions with the same composition, which is problematic during analysis using MS techniques. In addition, the most commonly used miRNA detection techniques including RT-PCR and NGS require labeling, RT, gene amplification, or sequencing (Git et al. 2010; Wang et al. 2011). Given this, our findings concerning discrimination of single-base mismatches using the ToF-SIMS technique are promising and allow for the development of new biosensing platforms with a high degree of specificity.

We next examined the sensing ability of our technique to detect target miRNAs in biological samples. Since certain miRNAs are potent prognostic and diagnostic biomarkers in various diseases, including cancers, the ability to detect miRNAs in cancer cell extracts would be a powerful weapon in the fight against the disease (Deng et al. 2017; Patnaik et al 2017; Zhang et al. 2017). However, the high concentration and diversity of RNAs within the total RNA extracts from cancer cells can interfere with hybridization of a low-abundant target. Thus, we first tested whether our sensor could detect a specific miRNA in total RNA extracts. To this end, we spiked the total RNA extracts from the SNU1 cells with miR-200a-3p at 1 nM and compared the signal obtained when using the same concentration of miR-200a-3p in buffer, using a gold chip modified with a hairpin DNA probe complementary to miR-200a-3p. No significant difference in the intensities of the peak at m/z 748.3 in the two samples was observed (Figure 3d), which indicates that the high concentration of diverse RNA molecules did not affect the interaction between the miRNA and the probe.

Based on this result, we further tested the ability of our sensing platform to detect the target miRNA in a real biological sample. We chose two kinds of human gastric adenocarcinoma cell lines: MKN45, which expresses miR-200a-3p; and SNU1, which does not. It is known that miR-200a is a key modulator to suppress epithelial to mesenchymal transition (EMT) process in various cancers (Davalos et al. 2012). The expression of miR-200a is also known to be associated with molecular subtype in gastric cancer (Song et al. 2014). Among molecular subtypes of gastric cancer, stem-like subtype is highly associated with EMT process and shows frequent recurrence and poor prognosis (Cristescu et al. 2015). In Figure S3, the expression of miR-200a was significantly down-regulated in the stem-like subtype in TCGA STAD cohort (t-test P < 0.001).

Based on the clinical importance of miRNA-200a in gastric cancer, we measured the expression levels in two different types of human primary gastric cancer cells (MKN45 and SNU1) using the EASE-ToF method. As shown in Figure 3e, the NBT-formazan molecule was detected in the total RNA extract from MKN45 cells, while no significant peak was observed in the total RNA extract from the SNU1 cells. The results obtained with our sensor were consistent with the results of RT-PCR (Figure 3e). Therefore, our sensing platform shows great potential in that low-abundant biomarkers can be detected in real biological samples, which is difficult to achieve using traditional approaches (Zhou et al. 2017); this feature is crucial for clinically-relevant sensing platforms.

Next, we investigated the potential of our analytical platform for multiplex assays utilizing the advantages of MS and well-known enzymes HRP, AP, and β-gal and their various substrates listed in Table S2. The product of the HRP reaction with the substrate AEC was observed at m/z 608.3, corresponding to [M-NH₃]⁺ (Figure 2a, see Figure S4a for the full spectrum); the product of the AP reaction with the BICP/NBT substrate was observed at m/z 748.3, corresponding to [M]⁻ (Figure 2b, see Figure S2b for the full spectrum) as mentioned above; and the product of β-gal reaction with the substrate S-gal was observed at m/z 231.1, corresponding to the deprotonated ion, [M-H]⁻ (Figure 2c, see Figure S4b for the full spectrum).

Furthermore, the spectrum (Figure S5) obtained from treatment with the three enzymes strongly suggests that our strategy could be applied to a multiplex assay. In addition, by using ToF-SIMS, not only the enzymatic reaction products of different enzymes could be detected, but also the products of the same enzyme reacting with different substrates produced distinguishable peaks. Specifically, we obtained the ToF-SIMS spectra of the products of the HRP reaction with the substrates 4-chloronaphthol (4-CN), AEC, and 3,3′-diaminobenzidine (DAB) (Figure S6). In addition, as shown in Figure S7, the peaks corresponding to products of β-gal reaction with the substrates 5-bromo-4-chloro-3-indolyl β-D-galactopyranoside (X-gal), S-gal, and 6-chloro-3-indolyl-β-D-galactopyranoside (Red-gal) were distinguishable, including in the spectrum where three substrates were coexisted.

Distinctively patterned chemical images corresponding to [M-NH₃]⁺ and [M]⁺ of enzymatic reaction products of AEC and BCIP/NBT, appeared at m/z 608.3 and 748.3, respectively, on the functionalized gold surface with excellent uniformity over a large area in Figure 4a. In addition, an image of [Au]⁺ at m/z 197.0 was observed surrounding the pattern in Figure 4b. In Figure 4c and 4d, average and standard deviation of intensities of five peaks were 1.2×10⁴ and 324 (2.6% deviation), respectively. Significantly, we obtained multiplexed images corresponding to six different peaks related to the enzymatic reaction products of X-gal, S-gal, Red-gal, AEC, and 3-hydroxy-2-naphthoic acid 2,4-dimethylanilide phosphate/4-chloro-2-methylbenzene diazonium (HNDP/CMD) as well as the Au ion image (Figure 4e). Images for multiple ions with different masses could be simultaneously extracted from a single measurement, as shown in Figures 4e and Figure S8, demonstrating that our biosensing strategy based on ToF-SIMS measurement on a functionalized gold chip platform has the potential for high molecular specificity enabling multiplexed assays to be performed in high-throughput analyses. These results further demonstrate that MS-based analyses of enzyme reaction products are well suited to characterize the intrinsic properties of molecules without considering their optical properties, which limits the use of chromogenic enzymes in multiplex assays.(Krieg and Halbhuber, 2013; Liu et al. 2015; Speel et al. 1994). As chromogenic reaction products are known to be stable for a long periods of time, it is advantageous to apply these processes to a biosensor as opposed to fluorescence or luminescence-based enzymatic substrates. Interestingly, our ToF-SIMS image showed that the signal was still well maintained when re-measured 6 months after the first measurement, as shown in Figure S9.

Figure 4.

Detection of various MSEs produced by enzymes. a) and b) Images measured by ToF-SIMS and reconstructed by peaks at m/z 197.0 (blue), 608.3 (green), and 748.3 (red), corresponding to Au (from the gold chip), the product of HRP reaction with AEC, and the product of AP reaction with BCIP/NBT, respectively (Image size: 54×17 mm²). c) Positions of selected region-of-interest (ROI). d) ToF-SIMS spectra extracted from selected ROIs. e) ToF-SIMS imaging for the detection of multiple enzyme reaction products including X-gal, S-gal, Red-gal, AEC, and HNDP/CMD.

More importantly, our EASE-ToF method is applicable not only to ToF-SIMS analysis but also to MALDI-ToF and LDI-ToF analyses (Figure S10). These results show that our signal amplification method can be used as a platform technology for general application in surface-based MS.

Next, we examined the possibility of protein detection to demonstrate the potential for universal usage of our biomolecule detection strategy. For protein detection, we chose interferon gamma (IFNγ), which is usually analyzed for tuberculosis diagnosis. After immobilization of anti-IFNγ antibodies on a gold chip, the chip was incubated with IFNγ in the presence of 1% BSA, followed by treatment with a biotin-labeled anti-IFNγ antibody, further triggering the binding of the streptavidin-AP conjugate (Figure 5a). After treatment with BCIP/NBT, we detected the deposited NBT-formazan molecules by ToF-SIMS analysis using methods similar to the miRNA detection method described above. The intensity of the peak at m/z 748.3 corresponding to NBT-formazan, [M]⁻, was dependent on the IFNγ concentration (Figure 5b and c). Using a “sandwich” assay format, the target protein was specifically captured by antibodies on the gold chip. As in miRNA detection, the enzymatic reaction produced MSEs in a large excess over target proteins. These results confirm that our signal amplification strategy can be extended to the detection of proteins, and thus, the antibody-antigen interactions, in addition to the RNA-DNA interaction shown above, can also be transduced into strong mass signals.

Figure 5.

Detection of NBT-formazan produced by the enzymatic reaction of AP in the presence of IFNγ by ToF-SIMS. A) Scheme for the protein detection strategy, B) Spectra of NBT-formazan as a function of IFNγ concentrations, C) Relationship between the normalized intensities of peaks corresponding to [M]⁺ at m/z 748.3 and concentrations of IFNγ. Data points represent the mean ± SD.

Conclusion

In conclusion, we demonstrated that enzymatic reactions can produce insoluble MSEs-MS in a quantitative manner with high target specificity and reproducibility in a multiplexed analytical format on gold chips for ToF-SIMS analyses. To the best of our knowledge, this is the first study to utilize reporter enzymes and their insoluble enzymatic reaction products as a tool for signal amplification in a biosensor using MS techniques. This strategy enabled highly sensitive recognition of the target in biological samples without purification, labeling (or tagging) on target analyte, gene amplification, or sequencing. Our strategy allowed excellent sensitivity and sequence specificity that are difficult to achieve with traditional MS techniques. In addition, we have shown the possibility for the analysis of various MSEs by measuring and imaging with ToF-SIMS, confirming the potential for high molecular specificity and multiplexity in a high-throughput format, thus overcoming the limited application of substrates based on their optical properties.

This approach demonstrates the versatile applicability of MS techniques with commonly used reporter enzymes for the detection of target biomarkers in biological samples. In principle, there should be no limitation based on the molecular weight of the target, and the technique is applicable to the analysis of various interactions between biomolecules. We expect that this enzymatic reaction product-based detection of DNA-RNA and antigen-antibody interactions, using ToF-SIMS and (MA)LDI-ToF, can be employed to study protein-protein, protein-substrate, and RNA-protein interactions and be effectively implemented to study a broad range of biological phenomena using surface MS techniques. However, although our strategy has proven that MS analysis can be a useful analytical method for bio-sensing, further research is needed for the integration of portable and compact equipment and detection strategies.