Nanoplasmon-enhanced drop-screen for high throughput single-cell nucleocytoplasmic miRNA profiling

Jia Liu ab, Guoyun Sun a, Shih-Chung Wei a, Song Guo a, Weikang Nicholas Lin a and Chia-Hung Chen *c
aDepartment of Biomedical Engineering, National University of Singapore, 4 Engineering Drive 3, #04-08, 117583 Singapore
bSchool of Chemistry and Chemical Engineering, Nanjing University, 163 Xianlin Avenue, Nanjing, 210023 China
cDepartment of Biomedical Engineering, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon Tong, Hong Kong SAR, China. E-mail:

Received 12th December 2019 , Accepted 12th April 2020

First published on 14th April 2020

Cell nucleocytoplasmic profiles of microRNAs (miRNAs) are critical to determining a single cell's essential functionalities, such as cellular transcription, nucleus export and degradation, which gives a comprehensive view of cellular processes. Despite the importance of addressing nucleocytoplasmic heterogeneity, the challenge of high-throughput screening remains. Although a droplet-based approach was developed for single-cell miRNA assays, the challenge of quantifying miRNA with high sensitivity to indicate nucleocytoplasmic heterogeneity remains. In this study, a nanoplasmon-enhanced droplet screening platform was developed to quantify single-cell nucleocytoplasmic heterogeneity with the high sensitivity of 0.1 nM. Droplet screening and multiplexed plasmonic assays are synergistic: droplet screening is used to isolate single cells for high-throughput screening, while enhanced nanoplasmonic assays are conducted to precisely determine different types of miRNAs, addressing the cell nucleocytoplasmic profile. Here, two nucleic acid-functionalized plasmonic nanosensors, silver nanoparticles functionalized with designed sequences to target miRNAs, are synthesized. After the targets are bound, competitive formation of sensor-target hybrids interferes with plasmonic coupling between the nanoparticles, decreasing a fluorescence signal and thus enabling high-sensitivity single-cell miRNA quantification. Using the fluorescence signal change as a readout allows continuous-flow measurement to provide a single-cell nucleocytoplasmic profile in a high-throughput manner (∼100 cells per minute) for effective quantitative cell biology.


MicroRNAs (miRNAs) constitute a phylogenetically extensive family of small noncoding RNAs that are ∼22 nucleotides long and regulate many important biological processes, including cell fate determination, cell proliferation and cell death.1–3 Most miRNAs are distributed diffusely in the cytoplasm to regulate biological processes, while a fraction of miRNAs are localized in the cellular nucleus to exert regulatory functions at the transcriptional level.4–9 Nucleocytoplasmic miRNA heterogeneity is critical to determine the regulatory network at the transcriptional level in individual cells.

Several platforms have been developed to measure miRNAs in single cells. For example, the quantitative reverse transcription polymerase chain reaction (RT-qPCR) method was developed to address gene expression heterogeneity with dynamic information.10 RT-qPCR was conducted to assess nucleocytoplasmic miRNA heterogeneity but performing RT-qPCR requires subcellular fractionation and tedious enzymatic amplification, causing false positives. Deep sequencing of the subcellular distribution of miRNAs was investigated via next-generation sequencing technology.11,12 Western blot technology was improved to indicate miRNA nucleus import.13,14 Optical approaches, such as live-cell surface-enhanced Raman spectroscopy (SERS) and atomic force microscopy (AFM)-based technology,15–17 were investigated to analyze single-cell RNAs with subcellular resolution. Both SERS and AFM require the insertion of nano/microfibers or probes similar to those in endoscopy to sample cytoplasmic or nucleoplasmic fractions within individual cells for analysis. These approaches require one cell per experimental run and so are subject to the throughput limitations for statistical sample profiling. Flow cytometry is useful for high throughput single cell assay, while it reaches the limitation to obtain statistical single cell miRNA data with high resolution.18

Recently, droplet-based microfluidics has emerged as a high-throughput approach for single-cell screening. Cells and biological reactants are dispersed into an oil phase to form isolated water-in-oil (w/o) droplets for single-cell encapsulations for measurements. In 2007, Huebner A. et al.19 demonstrated single-cell protein expression measurements using droplet-based microfluidics. Subsequently, Clausell-Tormos J. et al.20 incubated mammalian cells and multicellular organisms within droplets for high-throughput screening. In 2015, a droplet single-cell RNA assay was developed for comprehensive biosample analysis.21 After cell lysis in a droplet, RNA targets were released into the droplet, where they initiated polymerization involving reverse transcriptase. Through the incorporation of barcoding hydrogel beads in the droplets to label the RNA of individual cells, comprehensive single-cell RNA information could be deconvoluted using PCR. Since then, a series of creative works has led to the development of various single-cell RNA/DNA/protein assays using droplet microfluidics, enabling applications in quantitative biology and bacterial detection. Droplet-based hybridization was investigated for high-throughput single-cell miRNA measurements.22 Hairpin RNA sensors were developed to show increased fluorescence signals when target miRNAs cleave hairpin tangles. However, a high background signal caused by nonspecific binding of hairpin RNA reached the measurement sensitivity limitation (∼1 nM) for miRNA heterogeneity determination.

In this study, a plasmon-enhanced drop-screen system using nanosensors was developed for high-sensitivity subcellular assays to determine nucleocytoplasmic miRNA heterogeneity in a high-throughput manner (Fig. 1). The devices to produce the droplets for cell encapsulations and single cell screen are fabricated by using photolithographic technology (ESI 1, Fig. S1). The integration of plasmonic nanosensors and droplet screening technology is synergistic. Droplet-based encapsulation is used to ensure single-cell isolation for measurement, while plasmon-enhanced nanosensors are used to precisely identify different types of nucleocytoplasmic miRNAs. Two nucleic acid-functionalized nanosensors are synthesized: silver nanoparticles (AgNPs) modified with a sequence complementary to the target miRNA sequences (capture nanosensor) and AgNPs functionalized with a fluorophore-labeled sequence (indicator nanosensor). Capture nanosensors and indicator nanosensors are reacted to form sensor-sensor complexes to show high fluorescence signals resulting from a strong electromagnetic field produced by the plasmonic coupling between two adjacent plasmonic nanoparticles. Triton X-100, sodium dodecyl sulfate (SDS), tris-EDTA buffer solution (pH 8.0) and NaCl were used to prepare the nucleocytoplasmic cell lysis buffer for lysing the cells within the droplets (ESI 2, Table S1). Once target miRNAs bind nanoparticles, the sensor-target hybrids are formed by competitive binding between miRNAs and capture nanosensors to interfere with the magnetic field, resulting in significantly decreased fluorescence. The decrease in the fluorescence signal reflecting the binding events between miRNAs and nanosensors prevents background fluorescence noise to enable high-sensitivity measurements (∼0.1 nM). The single cell miRNA concentration is quantified by measuring the fluorescence signals in a single droplet. The drop-screen technology is developed to quantify target single cell miRNA concentration in one droplet. Droplets in a continuous flow are detected by a dark-field microscope (1000 droplets per second) for high-throughput screening. ∼100 single cells are measured individually (within the individual droplets) to obtain statistical data for biological sample profiling. The human mammary cancer cell line MCF-7 was cultured for testing. With the unique advantages of high-throughput single-cell screening with enhanced sensitivity, statistically effective of cytoplasmic miR-155 and nucleus miR-25 in the MCF-7 cell line are performed.

image file: c9lc01226e-f1.tif
Fig. 1 Schematic of enhanced nanoplasmonic droplet assay to determine nucleocytoplasmic miRNA heterogeneity in single cells. Once target miRNAs from the cell nucleus and cytoplasm bind to the plasmonic nanosensors, the paired nanoparticles are forced to separate, causing a dramatically enhanced plasmonic signal decrease to precisely quantify target miRNA concentrations within a single cell's nucleus and cytoplasm.

Materials and method

Plasmonic nanosensor fabrication

Silver nanoparticles (AgNPs) are fabricated using a citrate-mediated reduction reaction.23 In the first step, AgNO3 (36 mg) is dissolved in 200 mL of water by continuous stirring. Then, 4 mL of 1% (w/v) trisodium citrate is added to the solution. The mixture is boiled with stirring for ∼1 hour and then cools to 25 °C (room temperature). In the second step, to fabricate nucleic acid-functionalized AgNPs, the paired capturer and indicator are fabricated by modification of silver nanoparticles. Capture nanosensors are prepared by conjugating a complementary strand (capture sequence) to silver nanoparticles to identify the target miRNAs. Indicator nanosensors are prepared by conjugating a fluorophore-labeled sequence (indicator sequence) to AgNPs to show enhanced plasmonic fluorescence when they bind paired capture nanoparticles. A cytoplasm capture sequence (C-capture sequence, 10 μM, 2 μL) is mixed with AgNPs (18 μL) in an aqueous solution to fabricate cytoplasm-miRNA nanosensors. The same procedure is applied to prepare nucleus-miRNA nanosensors. A nucleus capture sequence (N-capture sequence 10 μM, 2 μL) is mixed with AgNPs (18 μL) in an aqueous solution to fabricate nucleus-miRNA nanosensors.

The synthesis of plasmonic nanosensor includes two parts: 1) AgNPs synthesis and 2) nucleic acid functionalization forming AgNP couple. The sequences are attached to the surface of the AgNPs by Ag–S bond-mediated self-assembly. The detailed sequence information is attached (ESI 3, Table S2). After a 2-hour reaction at room temperature, the supernatant is discarded by centrifugation at 6000 rpm for 5 min. The sediments, consisting of nanosensors, are resuspended in 20 μL of water and stored at 4 °C for the experiments. Two representative fluorescent dyes, Cy3 and Cy5, are employed as indicators of cytoplasmic and nucleus miRNAs, respectively.

Electromagnetic field simulation

The electromagnetic near-field enhancement of AgNPs are simulated by using a computer software, finite-difference time-domain method (FDTD, Version 8.6, Lumerical, Vancouver, Canada) based on Maxwell's equations. AgNPs with diameters ∼60 nm are used as the examples. The refractive index is provided by a database of Palik (provided by FDTD Solutions, Lumerical). The incident light is along with the y-axis. The polarization is along with the x-axis. A nonuniform mesh is used to simulate the region tested. The wavelength of incident light is set at 549 nm. Mesh step conditions of dx = dy = dz = 0.2 nm are employed. A standard Fourier transform is conducted to visualize the electromagnetic field distributions in the xy plane at the present wavelength.

Cell preparation

Cells of a human breast cancer cell line (MCF-7) was purchased from American Type Culture Collection (VA, USA). MCF-7 cells were cultured in Dulbecco's modified eagle medium (DMEM) supplemented with 10% fetal bovine serum (HyClone Laboratories, Logan, UT), penicillin (100 U mL−1; Life Technologies, Grand Island, NY) and streptomycin (100 μg mL−1; Life Technologies, Grand Island, NY). All cells were maintained in a humidified incubator at 5% CO2 until 90% confluency was reached and then collected for the evaluation of cell numbers. To evaluate the cell numbers, 10 μL of cell solution was mixed with 10 μL of 0.4% trypan blue stain (Life Technologies, USA). The mixed solution was then loaded onto a countess cell counting chamber slide (Invitrogen, USA). Cell viability was maintained above 90% for single-cell assays. After centrifugation at 200 g for 1 min, the supernatant was removed. The cells were resuspended in fresh DMEM to achieve a cell concentration of ∼ 5 × 106 cells per mL. Then, 15% OptiPrep density gradient medium (Sigma-Aldrich, USA) was added to prevent the sedimentation of cells in the syringe.

Drop screening system

An automatic optical system is developed for high-throughput droplet screening with a continuous flow manner (the throughput is ∼1000 droplets per second). Before screening occurs, droplets containing MCF-7 breast cancer cells and sensor-sensor complexes for miR-155 and miR-25 are incubated at 37 °C for 2 hours. After incubation, the droplets are reinjected into the device. Excitation light passes through a multiband bandpass excitation filter, and individual emission filters provide real-time readout of multiple fluorescent signals from two sets of photomultiplier tube (PMT) sensors. Excitation light passes through a multiband bandpass excitation filter, and individual emission filters enable a rapid readout of multiple fluorescent signals from two PMT sensors. A data acquisition (DAQ) system is applied to convert the analog voltage signals from the PMTs into digital signals for analysis. The flow rates of droplets and the spacing oil are set to 3 μL min−1 and 10 μL min−1, respectively.

Results and discussion

Characterization of plasmonic nanosensor

To quantify miRNAs from single cells with high sensitivity, nucleic acid-functionalized plasmonic silver nanosensors (AgNPs) are prepared (Fig. 2A). The paired nanoparticles (capture nanosensor and indicator nanosensor) formed AgNP couples as the plasmonic nanosensor (Fig. 2B). To evaluate the sequential binding to the AgNPs, UV-visible absorption spectra of the silver-based capture and indicator nanosensors for cytoplasmic miR-155 are characterized (Fig. 2C). Before binding occurs, a broad absorption peak at approximately 420 nm is observed for unmodified AgNPs (gray short-dashed line). A weak absorption peak at ∼260 nm is observed in the spectra of the C-capture sequence and the C-indicator sequence (dark dotted line and gray dotted line). After the capture sequence (or indicator sequence) is bound to the AgNPs, the spectra are similar to the spectra of unmodified AgNPs, showing a broad peak at approximately 420 nm (dark solid line and gray solid line). This result shows that the binding of the targeting miRNA does not change the optical properties of AgNPs used to measure the cytoplasmic miR-155 target via plasmonic enhancement changes. The stability of AgNP dimers is evaluated via UV-visible absorption spectrum. It is found that the absorbance at ∼420 nm is stable, regardless RNA dye binding on AgNP surface. The result indicates that AgNP dimers are formed successfully. AgNPs are not randomly aggregated after surface modification. Besides, a representative transmission electron microscopy (TEM) image of nucleic acid-functionalized AgNPs is recorded (Fig. 2C, insert figure), showing that their core diameters are ∼60 nm, and there are no obvious change in size after surface modification with 22-base nucleic acid sequences (for example, the miR-155 C-capture sequence is 5′-SH-CCCCUAUCACAAUUAGCAUUAA).
image file: c9lc01226e-f2.tif
Fig. 2 Synthesis and characterization of silver nanosensors for nucleocytoplasmic miRNA screening. (A) Schematic of the synthesis route for raw AgNPs to fabricate capture nanosensors and indicator nanosensors. (B) Schematic of capture sequences and indicator sequences attached to each AgNP. (C) UV-visible absorption spectra of raw AgNPs, silver-based capture and indicator nanosensors, capture and indicator sequences of miR-155, and capture-indicator complexes. Insert picture is a representative transmission electron microscopy (TEM) image of nucleic acid-functionalized AgNPs. (D) UV-visible absorption spectra of raw AgNPs, silver-based capture and indicator nanosensors, capture and indicator sequences of miR-25.

Similar results are observed in the case of plasmonic nanosensors for nucleus miR-25 (Fig. 2D). Without binding (unmodified AgNPs), the same broad peak at ∼420 nm is observed. In the case of N-capture sequences, two main peaks (∼260 nm; ∼650 nm) are observed. In the case of the N-indicator sequence, one weak main absorption peak at approximately 250 nm is observed. After binding occurs (both the N-capture nanosensor and N-indicator nanosensor), as observed in the optical properties of unmodified AgNPs, only one broad peak at approximately 420 nm is observed, showing that the optical properties of AgNPs are unchanged after N-capture sequences and N-indicator sequences are bound. The detailed nanosensor fabrication process is included in ESI 4. For comparison, gold-based nanosensors targeting the two representative miRNAs are also fabricated and characterized, as shown in Fig. S2. The experiments show that silver-based nanosensors generate higher fluorescence signal changes than gold-based nanosensors.

Fluorescence intensity enhancement

The mechanism to trigger plasmon-enhancement to measure the target miRNA is illustrated in Fig. 3A. Two major reactions are included: control binding and competitive binding. In the first step, the capture nanosensor and indicator nanosensor for each miRNA are mixed in a solution at 37 °C for 2 hours, forming the plasmon-enhanced nanosensors (AgNP couple). When paired, adjacent AgNP bind, based on a strong electromagnetic field produced by the plasmonic coupling effect, a high fluorescence signal is observed.24–27 To characterize the reactions for these measurements, the samples with miRNAs (miR-25, miR-155) are loaded. The mixtures are incubated at 37 °C for 2 hours. If the miRNA sequence loaded does not match the capture nanosensor sequence, the miRNA does not interact with the capture nanosensor, and there is no obvious change in fluorescence signal. When the miRNA sequence loaded matches the capture nanosensor sequence, the competitive binding between target miRNA and indicator nanosensors is established, allowing forming sensor-target hybrids, decreasing the fluorescence intensity.
image file: c9lc01226e-f3.tif
Fig. 3 Plasmonic measurement. (A) Schematic of the two binding modes in target miRNA recognition processes. (B) Electric field intensity simulated by FDTD in nontarget mode (sensor-sensor complexes) and competitive binding mode (sensor-target hybrids). (C) Fluorescence intensity of the two binding modes used to quantify miR-155 and miR-25. The concentration of miRNA is 100 nM. The buffer contains 50 mM NaCl (pH 7.6). The error bars are based on standard deviations of three parallel measurements. (D) The corresponding fluorescence emission spectra of two binding modes for detecting miR-155 (Ex/Em: 549/568 nm) and (E) miR-25 (Ex/Em: 648/668 nm), respectively.

The near-field enhancement properties of silver-based nanosensors are modeled by FDTD. The simulated electromagnetic field is illustrated in Fig. 3B. It is observed that the control binding mode (sensor-sensor complexes) generates a stronger electromagnetic field between the two AgNPs than the competitive binding mode (sensor-target hybrids). In the case of sensor-sensor complexes, the fluorophore is located in the middle between two AgNPs at a distance of ∼3–4 nm, showing the highest fluorescence signal due to the plasmonic coupling effect (ESI 5, Fig. S3). Once target miRNAs appear and competitively bind capture nanosensors to form sensor-target hybrids, the electromagnetic field significantly interferes, showing fluorescence decrease. The altered optical response recorded is a combination of attenuated fluorescence and scattering when the sensor-sensor complexes are cleaved. The level of fluorescence optical signal decrease is conducted to evaluate miRNA concentration. The detailed sensing mechanism of plasmonic fluorescence enhancement is elaborated in ESI 5.

The effectiveness of the signal enhancement of the control binding and competitive binding is experimentally verified. In the case of miR-155 detection, the capture nanosensor and indicator nanosensor react at 37 °C for 2 hours, showing an enhanced, strong fluorescence signal (Fig. 3C). In the presence of target miR-155, the targeted miRNA competitively binds capture nanosensors, resulting in a decrease in fluorescence from 3495 a.u. to 1390 a.u. In the case of miR-25, the fluorescence signal decreases from 1523 a.u. to 598 a.u. The corresponding fluorescence emission spectra of the two binding modes of miR-155 and miR-25 are shown in Fig. 3D and E, respectively.

Fluorescence signal characterization

To systematically characterize the fluorescence signals generated by plasmonic nanosensors, dynamic characterizations of two hybridization reactions are investigated as a function of time, as shown in Fig. 4A and B. The fluorescence signal is constant before target miRNAs bind. Once target miRNAs are loaded, the fluorescence signal decreases (2000 a.u. to 1500 a. u.) for 30 min to ∼1 hour. After 2 hours, a stable decrease in the fluorescence signal is observed. With this reference, a time period of 2 hours is set for our experiments.
image file: c9lc01226e-f4.tif
Fig. 4 Performance of the plasmonic nanosensors. (A and B) Time-dependent monitoring of the fluorescence intensity in control binding and competitive binding of (A) miR-155 and (B) miR-25. (C) Fluorescence emission spectra of an interference test for miR-155. Samples: 100 nM miR-155/miR-155A (single-base mismatch)/miR-155B (two-base mismatch)/miR-155C (three-base mismatch) in 20 mM Tris buffer containing 50 mM NaCl, pH 7.6. Inset data show the cross-reactivity for different mismatched miRNAs, and the error bars represent the standard deviations calculated for three parallel measurements. (D) Fluorescence emission spectra of the interference test for miR-25. Samples: 100 nM miR-25/miR-25A (single-base mismatch)/miR-25B (two-base mismatch)/miR-25C (three-base mismatch) in 20 mM tris buffer containing 50 mM NaCl, pH 7.6. Inset data show the cross-reactivity for different mismatched miRNAs, and the error bars represent the standard deviations calculated for three parallel measurements. (E) Fluorescence emission spectra to quantify miR-155. (F) Fluorescence signal changes upon control binding and competitive binding of varying concentrations of miR-155 (0.1 nM to 1000 nM). Inset figure shows the reduction in fluorescence intensity for miR-155 measurement at different concentrations. (G) Fluorescence emission spectra of the sensitivity test for miR-25. (H) Fluorescence intensity change of control binding and competitive binding for varying concentrations of miR-25 (0.1 nM to 1000 nM). Inset figure shows the reduction in fluorescence intensity for miR-25 measurement at different concentrations. The error bars included are based on standard deviations of three parallel measurements.

Selectivity is an important criterion for evaluation of the performance of plasmonic sensors. The selectivity to distinguish target miR-155 and miR-25 in mixtures is characterized in Fig. 4C and D, respectively. Without binding to target miRNAs, control binding for both miR-155 and miR-25 show high fluorescence intensity. With target cytoplasmic miR-155, the fluorescence intensity decreases from 3531 a.u. to 1390 a.u. due to the formation of sensor-target hybrids. In the case of nucleus miR-25, once the sensor binds to the target nucleus miR-25, the fluorescence intensity decreases from 1318 a.u. to 598 a.u. The negative control experiment using mismatched miRNAs shows limited fluorescence intensity decreases. For example, in the case of miR-155C, which has three bases mismatched from miR-155, the cross-reactivity between miRNAs and nanosensors is weak, causing a fluorescence decrease of ∼1.3% (Fig. 4C). A similar result is observed for the measurement of miR-25 (Fig. 4D). In different mismatch situations, different levels of fluorescence decreases are observed. Thus, plasmonic sensors fabricated in this work showed highly specific properties toward to target miRNAs.

Sensitivity is another key criterion for the evaluation of the plasmonic nanosensors. The sensitivity is characterized by monitoring the fluorescence intensity changes in solutions with different target miRNA concentrations. The concentrations of sequences used to prepare the capture nanosensor and indicator nanosensor are constant (10 μM). The concentration of AgNP used is about 3.6 × 10−11 M.28 The concentrations of target miRNAs are varied from 0.1 nM to 1000 nM. In solutions, the fluorescence signal distinctly decreases with increasing concentrations of the target miRNAs, suggesting that the plasmonic coupling effect in sensor-sensor complexes undergoes different degrees of interference at different concentrations of targets. For example, in the case of miR-155 measurement, when the miR-155 concentration increases from 0.1 nM to 1000 nM, the fluorescence decreases from 3495 a.u. to 789 a.u. (Fig. 4E and F). A linear relationship between reduced fluorescence intensity and target miRNA concentration is observed (inset data in Fig. 4F). Without the binding target miR-155, a high fluorescence signal is observed at ∼568 nm. In the presence of target miR-155, the intensity of this peak decreases. For example, at a concentration of 1000 nM, the fluorescence intensity is ∼789 a.u., while at a concentration of 0.1 nM, the fluorescence intensity is ∼3220 a.u. The low concentration (0.1 nM) of miR-155 is well determined by the fluorescence decrease (from 3495 a.u. to 3220 a.u.). The same fluorescence decrease trend is observed in the detection of miR-25. In this case, when the miR-25 concentration increases from 0.1 nM to 1000 nM, the fluorescence decreases from 1523 a.u. to 385 a.u. (Fig. 4G and H). Without the binding target miR-25, a strong emission peak is observed at ∼ 668 nm. In the presence of miR-25, the intensity of this peak decreases. A low concentration (0.1 nM) of miR-25 is determined by the fluorescence decrease (from 1523 a.u. to 385 a.u.).

Single-cell nucleocytoplasmic analysis

To validate the capability of the plasmon-enhanced droplet assay, high-throughput subcellular miRNA heterogeneity analysis is performed (Fig. 5A). The microfluidic device design used for cell encapsulation is shown in Fig. S1. The width of the microchannels is 40 μm. There are three inlets: 1) a cell suspension inlet, 2) a sensor solution inlet and 3) a carrier oil inlet. The flow rates of the oil and the two aqueous solutions are set to 10 μL min−1 and 3 μL min−1, respectively, forming water-in-oil droplets with diameters of ∼30 μm for single-cell encapsulation (Fig. 5B). With the advantage of high-sensitivity quantification of miRNAs, single-cell nucleocytoplasmic profiling is performed by quantifying cytoplasmic miR-155 and nucleus miR-25.29,30 After incubation of droplets with single cells and sensors at 37 °C for 2 hours, the droplets are loaded into a chamber for long-term observation and recording of the fluorescence intensity decrease (Fig. 5C). The cells are labelled by using Hoechst 33342 fluorescent stain, which generates a blue fluorescence signal (wavelength 460 nm). Cytoplasmic miR-155 is reported by Cy3, which shows a green fluorescence signal (wavelength 568 nm), while nucleus miR-25 is reported by Cy5, which shows a red fluorescence signal (wavelength 668 nm). Single-cell nucleocytoplasmic heterogeneity is determined by simultaneous measurements of the blue fluorescence signal (single cells), green fluorescence signal (miR-155 in the cellular cytoplasm) and red fluorescence signal (miR-25 in the cellular nucleus). The droplets containing single cells indicate different degrees of low fluorescence in both the green channel and the red channel, showing cell heterogeneity. Multiplexed single cell screening is performed to indicate different expression levels of cytoplasmic miR-155 and nucleus miR-25 in individual cells (ESI 6, Fig. S4), suggesting nucleocytoplasmic heterogeneity.
image file: c9lc01226e-f5.tif
Fig. 5 Single-cell nucleocytoplasmic miRNA heterogeneity screening. (A) Scheme of the online PMT analysis system for drop-screen of single cell analysis. (B) Individual cells encapsulated in single droplets. (C) The fluorescence images obtained from the blue channel (Ex/Em: 350/460 nm), green channel (Ex/Em: 549/568 nm) and red channel (Ex/Em: 648/668 nm) when single cells are lysed in the droplets. The scale bar is 30 μm. (D) Cell heterogeneity of miRNAs (MCF-7 cell line) is observed by recording fluorescence signals using the high-throughput system. (E) Two subcellular miRNAs (cytoplasmic miR-155 and nucleus miR-25) are monitored simultaneously to address cell nucleocytoplasmic heterogeneity. (F) The concentrations of miR-155 and miR-25 from single cells (MCF-7) are quantified by fluorescence intensity decrease in the droplets. In this plot, miRNA concentration in one single cell is presented by one spot. The average miRNA concentration is obtained by averaging single cell miRNA concentration. The number of single cells measured is 310.

The fluorescence intensity decreases to identify miR-25 and miR-155 in the droplets with cell encapsulation is observed (Fig. 5D). The fluorescence signals are recorded by two PMTs. The signal at a wavelength of 568 nm is used to recognize cytoplasmic miR-155, while the signal at a wavelength of 668 nm is used to identify nucleus miR-25. The control experiments to measure the cell encapsulated droplets and empty droplets were processed to validate the testing results (Fig. 5D). Simultaneously recording two fluorescence signals in individual cells in the droplets shows the nucleocytoplasmic miRNA distribution of single cells (Fig. 5E). The number of droplets measured to the results for control assay was 6770. The number of droplets measured to the results for droplets without cells was 3086, and the number of droplets measured to the results for droplets with single cell encapsulation was 310. Quantitative analysis of miR-155 and miR-25 in the MCF-7 cell line is shown in Fig. 5F, addressing cell heterogeneity in MCF-7 cell line (each spot represents a single cell's activity). The averaged miRNA concentration (box chart) indicated that both miR-155 and miR-25 are highly expressed in the MCF-7 cell line, consistent with results in previous studies.31–34


In summary, an enhanced plasmonic droplet assay is developed for sensitive and high-throughput single-cell subcellular miRNA screening. Based on the fluorescence signal decrease caused by target miRNA competitively binding to plasmonic nanosensors, cell nucleocytoplasmic heterogeneity is determined. The highest fluorescence signal is obtained when paired AgNPs are bound with each other through interactions of paired RNAs. Target miRNAs bind to the AgNPs due to their binding efficiency, forming sensor-target hybrids to separate the nanoparticles. This behaviour changes the plasmonic enhancement to cause a decrease in the fluorescence signal, allowing target miRNA quantification. With droplet encapsulation, miRNAs from the nucleus and cytoplasm of individual cells are determined by measurement of the fluorescence signal decrease. miRNA concentration from one single cell is quantified within one droplet. ∼100 single cells are measured individually within individual droplets to obtain statistical data. It is worth noting that it is desirable to measure more than ∼100 single cells for a biological sample profiling with statistical meaning. The miRNAs from the cytoplasm (miR-155) and miRNAs from the nucleus (miR-25) in MCF-7 cells are simultaneously measured with a throughput of ∼100 cells per second. With the advantage of a low fluorescence background, the detection sensitivity approaches ∼0.1 nM. The capability of using plasmon-enhanced droplet assays to rapidly profile large-scale single-cell nucleocytoplasmic heterogeneity of miRNAs allows precise investigation of biological samples to provide insights into comprehensive biological processes at the subcellular level.

Conflicts of interest

There are no conflicts to declare.


We gratefully acknowledge the funding provided by the City University of Hong Kong (9610467), National Research Foundation Singapore (R-397-000-276-281 and R-397-000-323-592), Synthetic Biology Research Program (NRF, SBP), National Research Foundation Singapore, Competitive Research Programme (NRF, CRP), National Medical Research Council Singapore (R-397-000-289-213), Open Fund – Individual Research Grant (NMRC, OFIRG) and Ministry of Education (MOE) Singapore (R-397-000-271-112 and R-279-000-501-112), Tier-2.


  1. L. He and G. J. Hannon, Nat. Rev. Genet., 2004, 5, 522–531 CrossRef CAS PubMed.
  2. H. F. Dong, J. P. Lei, L. Ding, Y. Q. Wen, H. X. Ju and X. J. Zhang, Chem. Rev., 2013, 113, 6207–6233 CrossRef CAS PubMed.
  3. J. Krol, I. Loedige and W. Filipowicz, Nat. Rev. Genet., 2010, 11, 597–610 CrossRef CAS PubMed.
  4. T. C. Roberts, Mol. Ther.--Nucleic Acids, 2014, 3, e188 CrossRef CAS PubMed.
  5. H. Y. Liu, C. Lei, Q. He, Z. Pan, D. S. Xiao and Y. G. Tao, Mol. Cancer, 2018, 17, 1–14 CrossRef PubMed.
  6. J. Y. Liao, L. M. Ma, Y. H. Guo, Y. C. Zhang, H. Zhou, P. Shao, Y. Q. Chen and L. H. Qu, PLoS One, 2010, 5, e10563 CrossRef PubMed.
  7. H. W. Liang, J. F. Zhang, K. Zen, C. Y. Zhang and X. Chen, Protein Cell, 2013, 4, 325–330 CrossRef CAS PubMed.
  8. D. H. Kim, P. Saetrom, O. Snove and J. J. Rossi, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 16230–16235 CrossRef CAS PubMed.
  9. J. Winter, S. Jung, S. Keller, R. I. Gregory and S. Diedeichs, Nat. Cell Biol., 2009, 11, 228–234 CrossRef CAS PubMed.
  10. A. K. White, M. Vanlnsberghe, O. I. Petriv, M. Hamidi, D. Sikorski, M. A. Marra, J. Piret, S. Aparicio and C. L. Hansen, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 13999–14004 CrossRef CAS PubMed.
  11. J. Y. Liao, L. M. Ma, Y. H. Guo, Y. C. Zhang, H. Zhou, P. Shao, Y. Q. Chen and L. H. Qu, PLoS One, 2010, 5, e10563 CrossRef PubMed.
  12. M. N. Abdelmoez, K. Lida, Y. Oguchi, H. Nishikii, R. Yokokawa, H. Kotera, S. Uemura, J. G. Santiago and H. Shintaku, Genome Biol., 2018, 19, 66 CrossRef.
  13. K. A. Yamauchi and A. E. Herr, Microsyst. Nanoeng., 2017, 3, 16079 CrossRef CAS PubMed.
  14. H. W. Hwang, E. A. Wentzel and J. T. Mendell, Science, 2007, 315, 97–100 CrossRef CAS PubMed.
  15. G. Lu, H. De Keersmaecker, L. Su, B. Kenens, S. Rocha, E. Fron, C. Chen, P. Van Dorpe, H. Mizuno, J. Hofkens, J. A. Hutchison and H. Uji-I, Adv. Mater., 2014, 26, 5124–5128 CrossRef CAS PubMed.
  16. J. Liu, Y. R. Wen, H. He, H. Y. Chen and Z. Liu, Chem. Sci., 2018, 9, 7241–7246 RSC.
  17. O. Guillaume-Gentil, R. V. Grindberg, R. Kooger, L. Dorwling-Carter, V. Martinez, D. Ossola, M. Pilhofer, T. Zambelli and J. A. Vorholt, Cell, 2016, 166, 506–516 CrossRef CAS PubMed.
  18. N. Nitta, T. Sugimura, A. Isozaki, S. Uemura, Y. Ozeki and K. Goda, et al. , Cell, 2018, 175, 266–276 CrossRef CAS PubMed.
  19. A. Huebner, M. Srisa-Art, D. Holt, C. Abell, F. Hollfelder, A. J. Demello and J. B. Edel, Chem. Commun., 2007, 1218–1220 RSC.
  20. J. Clausell-Tormos, D. Lieber, J. C. Baret, A. El-Harrak, O. J. Miller, L. Frenz, J. Blouwolff, K. J. Humphry, S. Koster, H. Duan, C. Holtze, D. A. Weitz, A. D. Griffiths and C. A. Merten, Chem. Biol., 2008, 15, 427–437 CrossRef CAS PubMed.
  21. A. E. Fischer, S. K. Wu, J. B. G. Proescher, A. Rotem, C. B. Chang, H. D. Zhang, Y. Tao, T. S. Mehoke, P. M. Thielen, A. O. Kolawole, T. J. Smith, C. E. Wobus, D. A. Weitz, J. S. Lin, A. B. Feldman and J. T. Wolfe, J. Virol. Methods, 2015, 213, 111–117 CrossRef CAS PubMed.
  22. S. Guo, W. N. Lin, Y. W. Hu, G. Y. Sun, D. T. Phan and C. H. Chen, Lab Chip, 2018, 18, 1914–1920 RSC.
  23. P. C. Lee and D. Meisel, J. Phys. Chem., 1982, 86, 3391–3395 CrossRef CAS.
  24. N. Gandra, C. Portz, L. M. Tian, R. Tang, B. G. Xu, S. Achilefu and S. Singamaneni, Angew. Chem., Int. Ed., 2014, 53, 866–870 CrossRef CAS PubMed.
  25. L. Touahir, E. Galopin, R. Boukherroub, A. C. Gouget-Laemmel, J. N. Chazalviel, F. Ozanam and S. Szunerits, Biosens. Bioelectron., 2010, 25, 2579–2585 CrossRef CAS PubMed.
  26. F. F. Yin, H. Q. Liu, Q. Li, X. Gao, Y. M. Yin and D. B. Liu, Anal. Chem., 2016, 88, 4600–4604 CrossRef CAS PubMed.
  27. M. H. Chowdhury, J. Pond, S. K. Gray and J. R. Lakowicz, J. Phys. Chem. C, 2008, 112, 11236–11249 CrossRef CAS PubMed.
  28. X. S. Bi, X. Z. Du, J. J. Jiang and X. Huang, Anal. Chem., 2015, 87, 2016–2021 CrossRef CAS PubMed.
  29. H. Q. Wang, X. D. Yu, Z. H. Liu, X. Cheng, D. Samartzis, L. T. Jia, S. X. Wu, J. Huang, J. Chen and Z. J. Luo, J. Pathol., 2011, 225, 232–242 CrossRef CAS PubMed.
  30. S. A. Khudayberdiev, F. Zampa, M. Rajman and G. Schratt, Front. Mol. Neurosci., 2013, 6, 43 Search PubMed.
  31. A. Esquela-Kerscher and F. J. Slack, Nat. Rev. Cancer, 2006, 6, 259–269 CrossRef CAS PubMed.
  32. Y. Wang and C. G. L. Lee, J. Cell. Mol. Med., 2009, 13, 12–23 CrossRef PubMed.
  33. H. Y. Zhang, Z. Zuo, X. Lu, L. Wang, H. Y. Wang and Z. L. Zhu, Oncol. Rep., 2012, 27, 594–598 CAS.
  34. Z. Y. Wang, N. Wang, P. X. Liu, Q. J. Chen, H. L. Situ, T. Xie, J. X. Zhang, C. Peng, Y. Lin and J. P. Chen, Oncotarget, 2014, 5, 7013–7026 CAS.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c9lc01226e

This journal is © The Royal Society of Chemistry 2020