Single-Molecule Sensor for High-Confidence Detection of miRNA

Abstract

MicroRNAs (miRNAs) play a crucial role in regulating gene expression and have been linked to many diseases. Therefore, sensitive and accurate detection of disease-linked miRNAs is vital to the emerging revolution in early diagnosis of diseases. While the detection of miRNAs is a challenge due to their intrinsic properties such as small size, high sequence similarity among miRNAs and low abundance in biological fluids, the majority of miRNA-detection strategies involve either target/signal amplification or involve complex sensing designs. In this study, we have developed and tested a DNA-based fluorescence resonance energy transfer (FRET) sensor that enables ultrasensitive detection of a miRNA biomarker (miRNA-342-3p) expressed by triple-negative breast cancer (TNBC) cells. The sensor shows a relatively low FRET state in the absence of a target but it undergoes continuous FRET transitions between low- and high-FRET states in the presence of the target. The sensor is highly specific, has a detection limit down to low femtomolar (fM) without having to amplify the target, and has a large dynamic range (3 orders of magnitude) extending to 300 000 fM. Using this strategy, we demonstrated that the sensor allows detection of miRNA-342-3p in the miRNA-extracts from cancer cell lines and TNBC patient-derived xenografts. Given the simple-to-design hybridization-based detection, the sensing platform developed here can be used to detect a wide range of miRNAs enabling early diagnosis and screening of other genetic disorders.

Keywords: biomarkers; fluorescence resonance energy transfer (FRET); high-confidence; miRNA; single-molecule; triple negative breast cancer (TNBC).

Figure 1.. Sensor design and working principle.

The sensor is comprised of synthetic single-stranded DNA, two of which are labeled with fluorophores to enable FRET. The sensor is expected to form a four-way DNA/RNA hybrid junction resulting in a rapid interconversion between low- and high-FRET states giving rise to dynamic FRET traces in single-molecule analysis.

Figure 2.. Typical single-molecule traces in the absence of target.

Intensity-time traces are shown on the left panel and the corresponding FRET traces are shown on the right for five representative molecules. In the absence of a target, the molecules showed a relatively static low FRET state. Experiments were carried out at room temperature (23 °C). FRET represents FRET efficiency (E_FRET).

Figure 3.. Typical single-molecule traces in the presence of miR-342-3p.

Intensity-time traces are shown on the left and the corresponding FRET traces are shown on the right for the five representative molecules. In the presence of miR-342-3p, the molecules showed clear dynamics between ~0.1 and ~0.5 FRET states. All experiments were carried out at room temperature (23°C). FRET represents FRET efficiency (E_FRET).

Figure 4.. Determination of the analytical sensitivity and specificity of the sensor.

A. Calibration curve showing the percentage of dynamic molecules against the target concentration. The inset displays the linear range of the calibration plot with a limit of detection (LOD) down to 50 fM. More than 150 single molecules were used at each concentration of target to calculate the percentage of dynamic molecules. The error bars represent standard deviation (SD) calculated using three randomly assigned groups of molecules at each concentration of miRNA. B. Test of sensor specificity using a near-saturating concentration (100 pM) of the miRNA target and a mutant. The sequence for the wild type and an A to C mutant are shown. C. Comparison of the sensor performance in 1× Tris buffer vs 10% human serum.

Figure 5:. RT-PCR analysis of miRNA extracts from cell lines and PDX samples of TNBC mice.

A. The fluorescence of the reporter dye (Rn) versus PCR cycle number (C_T) for synthetic miRNA-342-3p at various concentrations. Blank represents the sample processed through cDNA synthesis and real-time PCR without adding miRNA. The baseline of each curve is normalized to zero. B. Calibration curve for the miRNA plotted from the curves shown in part A. The C_T is plotted against the logarithmic concentration of miRNA and fitted using linear regression. C. Real-time PCR analysis of 5× diluted miRNA samples isolated from cancer cell lines and PDX tissues of TNBC mice using miRNA-342-3p–specific primers. Curves are average of two trials The baseline of the curves is normalized to zero. The determined concentrations of miRNA-342-3p in undiluted samples (extracts) are shown.