Isolation of DNA aptamers targeting N-cadherin and high-efficiency capture of circulating tumor cells by using dual aptamers

Tian Gao ab, Pi Ding ab, Wenjing Li b, Zhili Wang *b, Qiao Lin c and Renjun Pei *ab
aSchool of Nano-Tech and Nano-Bionics, University of Science and Technology of China, Hefei 230026, China
bCAS Key Laboratory for Nano-Bio Interface, Suzhou Institute of Nano-Tech and Nano-Bionics, Chinese Academy of Sciences, Suzhou, 215123, China
cDepartment of Mechanical Engineering, Columbia University, New York, NY 10027, USA

Received 27th August 2020 , Accepted 15th October 2020

First published on 17th October 2020


Abstract

Circulating tumor cells (CTCs) acquire mesenchymal markers (e.g., N-cadherin) and lose epithelial markers (e.g., epithelial cell adhesion molecule, EpCAM) during the epithelial-mesenchymal transition (EMT) and are therefore ideal biomarkers of tumor metastasis. However, it is still a challenge to efficiently capture and detect circulating tumor cells with different phenotypes simultaneously. In this work, to obtain aptamers targeting N-cadherin in the native conformation on live cells, we established stable N-cadherin overexpressing cells (N-cadherin cells) and used these cells to identify a panel of N-cadherin-specific aptamers through the cell-SELEX approach. Two aptamer candidates obtained after 12 rounds of selection showed a low equilibrium dissociation constant in the nanomolar range, indicating high binding affinity. The truncated aptamer candidate NC3S showed the highest binding affinity to N-cadherin cells with a low Kd value of 20.08 nM. The SYL3C aptamer was reported to target cancer cell surface biomarker EpCAM. Then, we synthesized two kinds of aptamer-modified magnetic nanoparticles (SYL3C-MNPs and NC3S-MNPs). Both SYL3C and NC3S aptamers possess excellent capture specificity and efficiency for the target cells. The aptamer–MNP cocktail exhibits a considerable capture efficiency and sensitivity for rare cancer cells of epithelial and mesenchymal phenotypes. Furthermore, no CTCs were found in blood samples from healthy donors, while CTCs were successfully isolated by using the aptamer–MNP cocktail for 15 out of 16 samples collected from patients. In summary, the two kinds of aptamer-modified MNPs could be utilized as a promising tool for capturing CTCs from clinical samples.


Introduction

Circulating tumor cells (CTCs) are cells that are shed from the primary tumor into the blood or lymphatic fluids and are responsible for distant metastasis. They are considered non-invasive “liquid biopsies”,1 which is of great significance for early diagnosis of tumors,2 prognosis monitoring,3,4 and understanding the metastasis mechanism.5 In addition, CTCs acquire mesenchymal markers (e.g., N-cadherin) and lose epithelial markers (e.g., epithelial cell adhesion molecule, EpCAM) during the epithelial-mesenchymal transition (EMT),6 and are therefore ideal biomarkers of tumor metastasis.7,8 Various techniques have been developed to separate and enrich CTCs, including immunomagnetic separation,9 nanostructure interfaces,10,11 microfluidic platforms,12,13 and physical properties (such as size/deformability, density).14,15 However, most of these methods are based on one single surface marker which will compromise the capture efficiency of CTCs. Furthermore, antibody-mediated CTC capture is limited by the fact that antibody–antigen binding is almost irreversible under normal physiological conditions, which makes rapid and non-destructive release of viable cells difficult.16 In addition, the antibody has several disadvantages, including variation between batches and high cost of production.17

The above limitations can be obviated by aptamers, also known as chemical antibodies, which offer the advantages of controlled variation between batches, easy synthesis and modification, low cost, etc.18,19 Aptamers are short single-stranded oligonucleotides that can fold into unique secondary or tertiary structures with high specificity and binding affinity to targets ranging from small molecules to whole cells.19,20 The cells captured by the aptamers can be completely released through nuclease hydrolysis, temperature-regulated conformational change, or a competing complementary sequence.16 The latter hybridizes with the aptamer, and effectively releases viable and functional cells from the aptamer surface without the need for any reagent.21 Therefore, dual aptamers against EpCAM and N-cadherin can be a promising strategy for capturing and purifying CTCs with epithelial and/or mesenchymal phenotypes.

Aptamers are generated in vitro by a selection process called systematic evolution of ligands by exponential enrichment (SELEX).22,23 Many aptamers were selected by using purified proteins as target molecules, but until now, useful aptamers are still very few. Since purified proteins differ from their native forms in terms of tertiary structure and function, the aptamers screened using purified proteins may not be able to recognize the same proteins expressed on the membrane of live cells.24 Compared with traditional selections that use purified proteins as the target, cell-SELEX, which uses whole live cells as selection targets, ensures that the generated aptamers can recognize the target cells. This obviates the need to purify proteins in their native conformations (a technically difficult process), as well as any prior knowledge of the target cell surface molecules.25,26 A series of aptamers against target cells was obtained with high affinity and specificity.27–29

In our previous work, a dual-antibody interface against EpCAM and N-cadherin was fabricated on the magnetic nanoparticles, exhibiting an improved capture efficiency of epithelial and mesenchymal target cells.16,30 In this study, we established the N-cadherin overexpressing cells (called N-cadherin cell),31 and used the cell-SELEX approach to generate aptamer candidates against N-cadherin. The control CHO-K1 and target N-cadherin cells were used for counter and positive selections. Through the living cell-SELEX,32 we have generated a group of DNA aptamers against N-cadherin that specifically recognized the N-cadherin cells. Among the aptamer sequences, the optimized aptamer NC3S with only 47-mer can bind to N-cadherin cells with a Kd value of 20.08 ± 0.72 nM. Then, we synthesized two kinds of aptamer-modified magnetic nanoparticles (SYL3C-MNPs and NC3S-MNPs) that target two types of cancer cell surface biomarkers EpCAM28 and N-cadherin respectively as previously described with modifications.30 The poly(carboxybetaine methacrylate) (pCBMA) was decorated on the surface of MNPs which showed little nonspecific cell adhesion. Both NC3S and SYL3C aptamers possess an excellent capture specificity and efficiency for the target cells. Simultaneously, the capture yields of the aptamer–MNP cocktail were observed to respectively be 92.5% for MCF-7 cells and 92.0% for HeLa cells, indicating a fine capture performance of the aptamer–MNP cocktail for epithelial and mesenchymal CTCs. In addition, the aptamer–MNP cocktail exhibited a considerable capture efficiency and sensitivity for rare cancer cells of epithelial and mesenchymal phenotypes. No CTCs were found in blood samples from healthy donors, while for 15 out of 16 samples collected from patients, CTCs were successfully isolated by using the aptamer–MNP cocktail. The capture specificity and efficiency for target cells, sensitivity for rare cancer cells, and capture performance for CTCs from blood samples of dual aptamers make them highly promising tools for the rapid detection and isolation of CTCs from clinical samples.

Materials and methods

CHO-K1 (Chinese hamster ovary cells), 293T (human embryonic kidney cells), KB (human oral epidermoid cells), K562 (human chronic myeloid leukemia cells), HeLa (human cervical cancer cells), Ca Ski (human cervical cancer cells), MCF-7 (human breast adenocarcinoma cells), H460 (human lung adenocarcinoma cells), GIST 882 (human gastrointestinal stromal tumor cells), and Hep G2 (human liver hepatoma cells) were obtained from the Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). CHO-K1 was maintained in F12K medium (GIBCO). KB, H460 and Ca Ski were cultured in RPMI medium 1640 (GIBCO). 293T, MCF-7 and GIST 882 were cultured in Dulbecco's minimal essential medium (DMEM, GIBCO). HeLa and Hep G2 were cultured in Eagle's minimum essential medium (GIBCO), K562 was cultured in Iscove's modified Dulbecco's medium (GBICO), and all media were supplemented with 10% FBS (GIBCO) and 100 U mL−1 penicillin–streptomycin. All cells were cultured in 100 mm × 20 mm or 60 mm × 15 mm culture dishes at 37 °C in a humidified 5% CO2 incubator. Cells were washed before and after incubation with washing buffer (4.5 g L−1 glucose and 5 mM MgCl2 in Dulbecco's PBS without calcium and magnesium). Binding buffer used for selection was prepared by adding yeast tRNA (0.1 mg mL−1, Sigma) and BSA (1 mg mL−1, Sigma) into washing buffer to reduce background binding.

Iron(III) chloride hexahydrate (FeCl3·6H2O), trisodium citrate, ethylene glycol (EG), ethanol, ammonium hydroxide, sodium acetate trihydrate (NaAc·3H2O), tetraethyl orthosilicate (TEOS), and N,N,N′,N′-tetramethylenediamine (TEMED) were purchased from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). Polyethylene glycol (Mw = 4000) was from Aladdin Co., Ltd (Shanghai, China). Ammonium persulfate (APS), N-hydroxysuccinimide (NHS), 3-(trimethoxylsilyl) propyl methacrylate (MPS), 1-ethyl-3-(3′-dimethylaminopropyl) carbodiimide (EDC), streptavidin (SA), Hoechst 33342, 3,3′-dioctadecyloxa-carbocyanine perchlorate (DiO), 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI), and Histopaque-1077 were purchased from Sigma-Aldrich. Alexa Fluor 555-labeled anti-Pan-Keratin (PanCK-555) and Alexa Fluor 488-labeled anti-CD45 (CD45-488) were purchased from Univ-bio Co. Ltd (Shanghai, China). 2-Carboxy-N,N-dimethyl-N-(2′-(methacryloyloxy)ethyl)-ethanaminium inner salt (CBMA) was synthesized according to our previous report.16 Deionized water (18.4 MΩ cm) used in all experiments was prepared using a Milli-Q system (Millipore, Boston, USA).

All oligonucleotides were obtained from Sangon Biotech Co., Ltd (Shanghai, China) and were high-performance liquid chromatography (HPLC)-purified by the manufacturer.

Initial ssDNA library (m-lib):

5′-ATACCAGCTTATTCAATT-40N-AGATAGTAAGTGCAATCT-3′

FAM-labeled forward primer: 5′-FAM-ATACCAGCTTATTCAATT-3′

Biotinylated reverse primer: 5′-biotin-AGATTGCACTTACTATCT-3′

Biotinylated SYL3C (biotin-SYL3C):

5′-Biotin-C6-CACTACAGAGGTTGCGTCTGTCCCACGTTGTCATGGGGGGTTGGCCTG-3′.

Plasmid construction

The full-length human N-cadherin/CD325/CDH2 ORF sequence was PCR amplified (20–30 cycles of 1 min at 95 °C, 1 min at 60 °C, and 15 min at 68 °C, followed by 10 min at 72 °C, and the rTaq-polymerase and dNTPs were obtained from TaKaRa). Then the PCR products were purified using a DNA gel extraction kit (Axygen), inserted into the pLVX-IRES-Puro plasmid following Xbal I (Takara) and BamH I (Takara) restriction, and ligated using T4 DNA ligase (Takara). E. coli HB101 electro-cells were transfected with N-cadherin-pLVX plasmids as per standard protocols, and the cloned insert was sequenced.

Establishment of N-cadherin overexpressing cells

To overexpress N-cadherin in CHO-K1 cells, a lentiviral vector mediated three-plasmid express system was used. Human kidney 293T cells were plated on 60 mm plates in 3 mL of culture medium and transfected after overnight incubation with 12 μg of N-cadherin-pLVX (or GFP-pLVX control), 8 μg of Pspax2, and 2 μg of pMD2.G premixed with Opti-MEM medium in the presence of Lipofectamine 2000 (Invitrogen). The supernatant was harvested 48 h post-transduction, cleared of cell debris by low-speed centrifugation, and filtered (0.45 μm). CHO-K1 cells were infected with the lentivirus-containing supernatant in F12K medium with polybrene (10 μg mL−1) (Solarbio). The virus was removed 12 h after transduction, and replaced with fresh F12K medium containing 1 μg mL−1 puromycin (Solarbio) to select the N-cadherin overexpressing cells. The overexpression of N-cadherin was assessed by flow cytometry. Briefly, the cells were incubated with polyclonal anti-N-cadherin antibody (Proteintech) or the polyclonal IgG isotype control (abcam) on ice for 30 min after blocking using 3% bovine serum albumin (Sigma Aldrich), and then with FITC-labeled secondary antibody (abcam) on ice for 20 min in the dark. After washing thrice with PBS, the cells were analyzed by flow cytometry.

Cell-SELEX procedures

The procedure of living cell-SELEX is as follows: the ssDNA library (10 nmol) dissolved in binding buffer was denatured at 95 °C for 5 min and cooled on ice for 10 min before incubation. Then the ssDNA library was incubated with N-cadherin cells (target cells) on ice for 1 h. After incubation, cells were washed thrice with washing buffer to remove unbound and weakly bound sequences. The adherent cells were scraped off and the bound DNA sequences were denatured at 95 °C for 10 min. The sequences were amplified by PCR using FAM and biotin-labeled primers (10–20 cycles of 0.5 min at 94 °C, 0.5 min at 58 °C, and 20 s at 72 °C, followed by 5 min at 72 °C; rTaq-polymerase and dNTPs were obtained from TaKaRa). After denaturing under alkaline conditions, the FAM-labeled ssDNAs were separated from the biotin-labeled antisense sequences using streptavidin-coated agarose beads, and purified by ethanol precipitation for the next round of selection.

The counter-SELEX with negative cells is usually performed to reduce non-specific binding.33 After three rounds of selection, the counter selection was performed with CHO-K1 cells (negative cell) to reduce non-specific binding. Briefly, the ssDNA library was incubated with the control CHO-K1 cells on ice for 30 min, and the unbound sequences were then incubated with N-cadherin cells. In order to acquire aptamers with high affinity and specificity, the selection pressure was gradually enhanced by decreasing the amount of the ssDNA pool, incubation time for the target cells, and the number of target cells, and by increasing the volume of washing buffer and the number of the counter selection cells. The enrichment of aptamers was monitored by flow cytometry, and after 12 rounds of selection, the enriched ssDNA library was amplified by PCR and cloned into BL21 E. coli using a PMD18-T TA cloning kit (TaKaRa). A total of 30 clones were randomly selected and sequenced by GENEWIZ Inc. (China). The sequences were subjected to sequence alignment analysis using Clustalx 1.8.3 software. Their secondary structures were predicted by means of the free-energy minimization algorithm using the internet tool M-fold (http://mfold.rna.albany.edu/).34

Flow cytometric analysis

In order to investigate the enrichment of the library, the cells were washed with PBS, detached with 0.02% EDTA dissociation solution, neutralized with culture medium containing 10% FBS and counted for further flow cytometric assay. Then, the FAM-labeled ssDNA library was incubated with 2 × 105 N-cadherin cells or CHO-K1 cells in 200 μl of binding buffer containing 10% FBS on ice for 50 min, washed thrice with 700 μl of washing buffer and re-suspended in 350 μl of washing buffer. The fluorescence intensity of the cells was analyzed by flow cytometry as a measure of aptamer enrichment by counting 15[thin space (1/6-em)]000 events. The FAM-labeled initial ssDNA library (m-lib) was used as the negative control.

To characterize the binding affinity of the aptamers, the N-cadherin cells (2 × 105) were incubated with varying concentrations of the FAM labeled-aptamer or FAM-labeled m-lib in 200 μL of binding buffer containing 10% FBS on ice for 50 min in the dark. Subsequent operations are described above. All the experiments for binding assay were repeated 2–4 times. The equilibrium dissociation constants (Kd) of the different aptamers were calculated using the following formula: image file: d0nr06180h-t1.tif. F and F0 represent the mean fluorescence intensity of the cells incubated with the FAM-labeled aptamer or m-lib; X represents the concentration of the FAM-labeled aptamer.

To test the specificity of aptamer candidates, the target cells and other cell lines were incubated with the three aptamers labeled by FAM, including CHO-K1, 293T, K562, MCF-7, KB, HeLa, Hep G2 and H460. Then the cells were analyzed by flow cytometry and the experiment was repeated thrice.

Confocal imaging of cells stained with aptamers

To further confirm that the enriched pool or aptamer candidates can bind to the target cells, the N-cadherin or CHO-K1 cells were cultured in a 35 mm glass bottom culture dish overnight, and then incubated with FAM-labeled m-lib (250 nM), an enriched pool or aptamer candidate (250 nM) in binding buffer containing 10% FBS on ice for 50 min in the dark. After washing with washing buffer thrice, the cells were imaged with an Olympus FV500-IX81 confocal microscope (Olympus America Inc., Melville, NY).

Effect of temperature on binding

Target N-cadherin cells or control CHO-K1 cells (2 × 105) were incubated with FAM-labeled aptamers at 4 °C, room temperature and 37 °C for 50 min in the dark, and the FAM-labeled m-lib was used as the control. Then the cells were subjected to flow cytometric analysis.

Synthesis of functionalized magnetic nanoparticles (MNPs)

Fe3O4 magnetic nanoparticles (MNPs) were prepared via a solvothermal reaction as previously reported with modifications.35 Briefly, 1.890 g of FeCl3·6H2O, 0.7 g of trisodium citrate, and 0.5 g of polyethylene glycol were dissolved in 70 mL of ethylene glycol. Then, 6.96 g of NaAc·3H2O was added and stirred vigorously for 1 h to form a homogeneous solution. The mixture was transferred into a Teflon-lined stainless-steel autoclave and heated to 200 °C and maintained for 10 h. MNPs were separated with the help of a magnet and washed with alcohol and deionized water, and then dispersed in deionized water for subsequent use. Next, the MNPs were modified with SiO2 and poly(carboxybetaine methacrylate) (pCBMA) to form the MNPs@SiO2 and MNPs@SiO2@CBMA according to our previous description.30 Specifically, 0.1 mg of MNPs was dispersed in 1 mL of ethanol–water solution (v[thin space (1/6-em)]:[thin space (1/6-em)]v = 4[thin space (1/6-em)]:[thin space (1/6-em)]1), and then 10 μL of TEOS and 20 μL of ammonium hydroxide were added into the solution under ultrasonication to form a silica shell on the surface of magnetic nanoparticles (MNPs@SiO2). After reaction, the MNPs@SiO2 products were collected using a magnetic scaffold for 5 min and washed with ethanol. Then, the MNPs@SiO2 were dispersed in ethanol solution containing 2% (v/v) MPS to form abundant double bonds on the surface of MNPs@SiO2. After that, the double bond-modified MNPs@SiO2 were reacted with 1% (w/v) CBMA solution with the help of 10 μL of APS (10%) and 10 μL of TEMED (5%) to obtain the MNPs@SiO2@CBMA. Finally, the MNPs@SiO2@CBMA was activated with 0.1 M of EDC and 0.025 M of NHS in 1 mL of MES solution, and then was immediately reacted with SA solution (20 μg mL−1) to immobilize SA. The SA-modified nanoparticles were incubated with biotinylated aptamers for 1 h (2 μM for biotin-SYL3C, or 2 μM for biotin-NC3S) to achieve aptamer-modified MNPs (SYL3C-MNPs, or NC3S-MNPs).

Cell capture experiments

The human breast cancer cell line (MCF-7, a high expression of EpCAM and a low expression of N-cadherin) and human cervical cancer cell line (HeLa, a high expression of N-cadherin and a low expression of EpCAM) were used as epithelial and mesenchymal subtype cancer cell models for investigating the capture behavior of functionalized MNPs, respectively. To investigate the capture efficiency of the functionalized MNPs, different concentrations of NC3S-MNPs were incubated with 105 of prestained HeLa cells in a 2 mL EP tube at 37 °C for 30 min. After cell capture, the cells captured by NC3S-MNPs were collected using the magnetic scaffold for 5 min and washed with PBS. The captured cells were imaged and counted with a fluorescence microscope to calculate the capture efficiency.

The cell isolation efficiency in artificial samples was tested to evaluate the sensitivity of the dual-aptamer system for capturing rare target cells. To obtain artificial samples, different numbers of prestained HeLa cells, MCF-7 cells, or the mixture of HeLa and MCF-7 cells (1[thin space (1/6-em)]:[thin space (1/6-em)]1) were suspended in 1 mL of PBS or white blood cell suspension (WBCs), respectively. The WBC suspension was separated from 1 mL of the fresh whole blood of healthy volunteers by centrifugation and Histopaque-1077 solution. After cell capture, the samples were washed with PBS, and the captured cells were counted using a fluorescence microscope to calculate the capture efficiency.

Capture of CTCs from clinical blood samples

The ethylenediaminetetraacetic acid anticoagulated blood samples were obtained from the Second Affiliated Hospital of Soochow University after obtaining ethical approval from the Ethics Committee of the Second Affiliated Hospital of Soochow University (EC-AF(JD)-06/6.1). Firstly, the blood samples were depleted of erythrocytes using Histopaque-1077 according to the manufacturer's instructions, and then incubated with the aptamer–MNP cocktail (ratio of SYL3C-MNPs to NC3S-MNPs of 1[thin space (1/6-em)]:[thin space (1/6-em)]1) for 30 min at 37 °C. After the cell isolation, the captured cells were fixed with 4% paraformaldehyde, and then blocked with 5% FBS and 0.3% Triton X-100 in PBS. After that, the samples were stained with Alexa Fluor 555-labeled anti-Pan-Keratin (PanCK-555) and Alexa Fluor 488-labeled anti-CD45 (CD45-488) in PBS at 4 °C overnight. Next, the samples were stained with Hoechst 33342 and imaged using a confocal laser scanning microscope. Cells that displayed PanCK+/CD45−/Hoechst 33342+ were morphologically intact and were identified as CTCs.

Results and discussion

The establishment of N-cadherin overexpressing cells

Before the SELEX, we examined the expression level of N-cadherin on transfected CHO-K1 cells by flow cytometry using anti-N-cadherin antibody. As shown in Fig. 1a, an increase in fluorescence intensity was observed in the presence of anti-N-cadherin antibody, but there was no significant change in the fluorescence intensity of CHO-K1 (Fig. 1b). The flow cytometry analysis results indicated that N-cadherin was overexpressed on the transfected CHO-K1 cells (called N-cadherin cell henceforth).
image file: d0nr06180h-f1.tif
Fig. 1 The expression of N-cadherin on transfected CHO-K1 cells (a) and native CHO-K1 cells (b). The cell-surface overexpression of N-cadherin was confirmed by flow cytometry using anti-N-cadherin antibody, and the IgG was used as the control.

293T cells that served as a packaging cell can produce a high titer recombinat lentivirus which can lead to stable transgene expression in target cells. Depending on the virus generation, three plasmids need to be premixed prior to adding to the complex to ensure that the transfection complexes are not formed preferentially with one plasmid over another.31 In order to establish a stable cell line, target cells were screened according to the resistance markers contained in the gene vector. So puromycin was used to select target cells expressing N-cadherin.

The generation of aptamers against N-cadherin by cell-SELEX

Since the molecular structure and morphology of cell surface proteins in the purified form might be different from those of the cell membrane, aptamers screened with purified proteins are often difficult to recognize the same proteins expressed on the membrane of live cells. In addition, cell-SELEX in particular involves direct interaction of the initial ssDNA pool with whole cells to generate aptamers that can detect the target cells or native surface proteins on cells.36 The method of using engineered cells is to introduce the gene of interest into cells for expression, and the natural conformation of the protein in the membrane can be maintained. In the screening of cell lines, the choice of control cells is particularly important. When the engineered cells are subjected to cell selection, the untreated cell lines are selected as the control, causing control conditions to be more stringent, which is more favorable for screening of aptamers for membrane proteins. The CHO-K1 and N-cadherin cells were used for negative and positive selection in this study. The process of our living cell-SELEX is shown in Fig. 2a. The initial ssDNA pool consisting of 40-mer random sequences flanked by two 18 nucleotide-long conserved sequences was used as an initial pool (m-lib). The N-cadherin cells were incubated with the initial pool and were scraped off after washing. The bound sequences were denatured by heating at 95 °C and then were enriched through PCR to generate an ssDNA pool for the next round of selection. In order to generate the ligands with high affinity and selectivity for N-cadherin, the negative cells were introduced to remove nonspecific sequences, and sequences unbound by negative cells were incubated with target cells for positive selection. When the number of positive selection rounds increased, the unbound DNAs were washed and the bound ssDNA pools were enriched.
image file: d0nr06180h-f2.tif
Fig. 2 Schematic illustration of the SELEX process (a) and the isolation and rapid identification of heterogeneous circulating tumor cells (CTCs) using the aptamer–magnetic nanoparticle cocktail (b).

After each round of selection, the selected FAM-labeled ssDNA pool was incubated with the N-cadherin or CHO-K1 cells, and the fluorescence intensity of cells was analyzed by flow cytometry. The FITC-H represented the binding capacity of enriched ssDNA pool to the cells. The fluorescence intensity of the target cells increased steadily with the number of selection rounds, indicating enrichment of the ssDNA library. No significant increase was observed in fluorescence intensity after the 12th round, which suggested that the N-cadherin specific ssDNA pool was saturated (Fig. S1a). In contrast, no obvious enhancement was seen in the fluorescence intensity of the negative control cells treated with the same (Fig. S1b). The N-cadherin binding of the selected pool was also confirmed by confocal imaging, as shown in Fig. S1c. Target N-cadherin cells and control CHO-K1 cells were incubated with FAM-labeled 12th pool or initial pool (m-lib) respectively. The N-cadherin cells treated with the 12th pool showed an intense fluorescence compared to being incubated with m-lib. The CHO-K1 cells displayed no detectable fluorescence signal result after incubation with the 12th pool or m-lib. The confocal imaging results are consistent with flow cytometry data, and they also suggested that the selected pool was enriched and recognized specifically the target cells.

The characterization of aptamer candidates

The enriched ssDNA pool was cloned and 30 clones were sequenced and grouped based on their homology (Fig. S2). Four selected aptamer sequences based on homology and possible secondary structure, including NC-2 and NC-3 from the family 2, NC-15 from the family 1 and NC-21, are shown in Table 1, and their putative secondary structures are shown in Fig. S3. The four aptamers were chosen for further characterization of the binding capacity to N-cadherin cells by flow cytometry.32 Compared with the FAM-labeled m-lib-treated group used as the negative control, the significant increase in fluorescence intensity of the N-cadherin cells was observed after incubating with the NC-2 and NC-3 sequences. In contrast, there was no obvious change in that of CHO-K1 cells (Fig. S4 and b). The flow cytometric assay results indicated that aptamer candidates NC-2 and NC-3 showed a relatively high binding ability to N-cadherin cells, but did not to the control CHO-K1 cells. However, as shown in Fig. S5, the flow cytometric results showed weak binding of NC-15 and NC-21.
Table 1 The detailed sequences of the selected aptamer candidates (primers are underlined)
Name Sequence (5′ to 3′)
NC-2 [A with combining low line][T with combining low line][A with combining low line][C with combining low line][C with combining low line][A with combining low line][G with combining low line][C with combining low line][T with combining low line][T with combining low line][A with combining low line][T with combining low line][T with combining low line][C with combining low line][A with combining low line][A with combining low line][T with combining low line][T with combining low line]AAGTAAGATTCCACTATGTTTTAGCTAGGGTTTCCTCCGG[A with combining low line][G with combining low line][A with combining low line][T with combining low line][A with combining low line][G with combining low line][T with combining low line][A with combining low line][A with combining low line][G with combining low line][T with combining low line][G with combining low line][C with combining low line][A with combining low line][A with combining low line][T with combining low line][C with combining low line][T with combining low line]
NC-3 [A with combining low line][T with combining low line][A with combining low line][C with combining low line][C with combining low line][A with combining low line][G with combining low line][C with combining low line][T with combining low line][T with combining low line][A with combining low line][T with combining low line][T with combining low line][C with combining low line][A with combining low line][A with combining low line][T with combining low line][T with combining low line]GAGTAAGAGTGCACTATGTTTTAGCTAGGGTTCCCTCCGG[A with combining low line][G with combining low line][A with combining low line][T with combining low line][A with combining low line][G with combining low line][T with combining low line][A with combining low line][A with combining low line][G with combining low line][T with combining low line][G with combining low line][C with combining low line][A with combining low line][A with combining low line][T with combining low line][C with combining low line][T with combining low line]
NC-15 [A with combining low line][T with combining low line][A with combining low line][C with combining low line][C with combining low line][A with combining low line][G with combining low line][C with combining low line][T with combining low line][T with combining low line][A with combining low line][T with combining low line][T with combining low line][C with combining low line][A with combining low line][A with combining low line][T with combining low line][T with combining low line]ACATGTCCTTAAGGGGAATAATATACAGCTTTGGGTGGTT[A with combining low line][G with combining low line][A with combining low line][T with combining low line][A with combining low line][G with combining low line][T with combining low line][A with combining low line][A with combining low line][G with combining low line][T with combining low line][G with combining low line][C with combining low line][A with combining low line][A with combining low line][T with combining low line][C with combining low line][T with combining low line]
NC-21 [A with combining low line][T with combining low line][A with combining low line][C with combining low line][C with combining low line][A with combining low line][G with combining low line][C with combining low line][T with combining low line][T with combining low line][A with combining low line][T with combining low line][T with combining low line][C with combining low line][A with combining low line][A with combining low line][T with combining low line][T with combining low line]GAAGAGTATAAAAAGAGTGATTATCTTTTGTAGGTTTTTT[A with combining low line][G with combining low line][A with combining low line][T with combining low line][A with combining low line][G with combining low line][T with combining low line][A with combining low line][A with combining low line][G with combining low line][T with combining low line][G with combining low line][C with combining low line][A with combining low line][A with combining low line][T with combining low line][C with combining low line][T with combining low line]
NC3S [T with combining low line][T with combining low line]GCACTATGTTTTAGCTAGGGTTCCCTCCGG[A with combining low line][G with combining low line][A with combining low line][T with combining low line][A with combining low line][G with combining low line][T with combining low line][A with combining low line][A with combining low line][G with combining low line][T with combining low line][G with combining low line][C with combining low line][A with combining low line][A with combining low line]


Then, the aptamer candidates NC-2 and NC-3 were chosen for further investigating the binding affinity to N-cadherin cells using flow cytometry. The different concentrations of FAM-labeled aptamers were incubated with N-cadherin cells for evaluation, whose equilibrium dissociation constants (Kd) were in the nanomolar range with a Kd value of 68.24 ± 13.04 nM and 62.76 ± 4.76 nM, respectively (Fig. S4c and d).

Optimization of the NC-3 aptamer

Generally, the selected full-length aptamer sequences do not always have satisfactory binding affinity for target molecules. Thus we further optimized that of NC-3 with the highest affinity to the N-cadherin cells based on its secondary structure as predicated by M-fold, and generated a truncated sequence that was designed as NC3S. As shown in Fig. 3, there was a clear right shift in the fluorescence intensity of target cells but not in negative cells (Fig. 3a and b), which in turn suggested that the forward primer is dispensable to the formation of the stem-loop structure required for binding to N-cadherin cells (Fig. 3c). And the NC3S did not lose binding capacity to the N-cadherin with a Kd value of 20.08 ± 0.72 nM (Fig. 3d).
image file: d0nr06180h-f3.tif
Fig. 3 Binding capacity of selected aptamer NC3S with N-cadherin and CHO-K1 cells assessed by flow cytometry and confocal imaging. Flow cytometry assays for the binding of NC3S with N-cadherin (a) and CHO-K1 cells (b). The predicted secondary structure of NC3S by M-fold (http://mfold.rna.albany.edu/) (c). Equilibrium dissociation constant (Kd) curve of NC3S for N-cadherin cells. The standard deviations were obtained from three separate trials (d). Confocal imaging of cells stained by NC3S (e). Images of CHO-K1 cells and N-cadherin cells after incubation with the FAM-labelled NC3S. In each picture, left is the optical image and right is the fluorescence image. The final concentration of the FAM-labeled sequence is 250 nM. The scale bar in the images is 50 μm.

The ability of aptamer NC3S to bind to N-cadherin cells was further confirmed by confocal fluorescence microscopy imaging. As shown in Fig. 3e, there was no obvious fluorescence signal with CHO-K1 incubated with m-lib or NC3S. However, a bright fluorescence signal was observed in N-cadherin cells stained with aptamer NC3S but not that stained with m-lib. The results clearly proved that aptamer NC3S displayed high binding ability toward N-cadherin cells.

Binding specificity of DNA aptamer candidates

The specificity of N-cadherin aptamer candidates NC-2, NC-3 and NC3S was evaluated using different cell lines. As shown in Table 2, all the aptamers recognized the N-cadherin cells but did not bind to CHO-K1 cells. Aptamer candidates also bound to HeLa, KB and H460 cells, which suggested that these aptamer candidates revealed their target binding affinity and specificity toward N-cadherin-positive cells.
Table 2 Binding assay of aptamers to different cell lines
Aptamer NC-2 NC-3 NC3S
A threshold of the fluorescence intensity of flow cytometry analysis was set so that 95% of cells incubated with the FAM-labeled ssDNA library would have fluorescence intensity below it. The percentage of the cells with fluorescence above the set threshold was used to evaluate the binding capacity of the aptamer to the cells. −, <10%; +, 11–35%; ++, 36–60%; +++, 61–85%; ++++ >85%.
CHO-K1
N-cadherin ++++ ++++ ++++
293T
H460 ++
MCF-7
Hep G2 + +
K562
HeLa ++ ++
Ca Ski +
KB + ++ ++
GIST 882


The effect of temperature on aptamer candidates

During the selection, cells were incubated with the ssDNA library at 4 °C, but further studies need to be carried out under physiological conditions. Therefore, the effect of temperature on their affinity to the N-cadherin cells was also analyzed. As shown in Fig. S6, we examined the binding capacity of NC-2, NC-3 and NC3S at different temperatures when incubating with target N-cadherin cells and control CHO-K1 cells. A greater shift in fluorescence intensity was observed at 37 °C compared to 4 °C, indicating that the aptamer candidates can bind to N-cadherin cells at physiological temperature. Taking these findings together, high affinity aptamers for N-cadherin were identified from a large pool of ssDNA using cell SELEX, which were highly specific for N-cadherin cells and were able to bind at physiological temperatures.

Capture performance of aptamer-MNPs in artificial samples

In our previous work, a fluorescent–magnetic nanoparticle platform with a dual-antibody interface against EpCAM and N-cadherin was developed, exhibiting a good capture efficiency and identification for epithelial CTCs as well as mesenchymal CTCs.30 Herein, an N-cadherin-specific aptamer through the cell-SELEX approach was obtained, displaying high binding affinity to N-cadherin cells. To confirm the ability of the N-cadherin-specific aptamer to selectively capture CTCs with high efficiency, the N-cadherin-specific aptamer was modified onto the surface of Fe3O4 magnetic nanoparticles (MNPs) according to our previous report,30 as shown in Fig. S7 and S8. MCF-7 cells and HeLa cells were employed as epithelial and mesenchymal subtype cancer cell models for cell capture experiments, respectively. The schematic illustration of the isolation and rapid identification of heterogeneous circulating tumor cells (CTCs) using the aptamer–magnetic nanoparticle cocktail is shown in Fig. 2b.

To assess the influence of nanoparticle concentration and incubation time on the capture efficiency for target cells, NC3S aptamer modified MNPs (NC3S-MNPs) and HeLa cells were first used to test the capture performance, as shown in Fig. 4a and b. It showed that the capture efficiency was improved with the increase of the nanoparticle concentration and incubation time, and the capture yied for HeLa target cells was approximately up to 90.6% using 0.1 mg mL−1 of NC3S-MNPs with an incubation time of 30 min. To confirm the effect of pCBMA brushes on nonspecific cell adhesion and aptamers on specific capture, cell experiments were performed on the differently modified MNPs, including MNPs, MNPs@SiO2@CBMA, different aptamer modified MNPs, and aptamer–MNP cocktails. The results are shown in Fig. 4c. Compared with bare MNPs, cell adhesion was distinctly reduced when pCBMA was modified onto the surface of MNPs. After modifying the different aptamers, the capture yied of NC3S-modified MNPs and SYL3C-modified MNPs for HeLa cells had a significant difference, which was due to the high expression of N-cadherin and low expression of EpCAM on the surface of HeLa cells. Additionally, the cocktail of SYL3C-MNPs and NC3S-MNPs at a ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]1 (aptamer–MNP cocktail) could excellently achieve the efficient capture of target cells. These results clearly indicate that pCBMA brushes have satisfactory ability of resisting nonspecific cell adhesion and the introduction of aptamers can successfully induce specific cell bingding to pCBMA-modified MNPs.


image file: d0nr06180h-f4.tif
Fig. 4 Optimization of the aptamer modified MNPs for cell capture using artificial samples. Capture efficiency of NCS3-modified MNPs for HeLa cells at different concentrations (a). Capture efficiency of NCS3-modified MNPs for HeLa cells at different incubation times (b). Comparison of capture efficiency for HeLa cells using the different interfacial modified MNPs (bare MNPs, pCBMA modified MNPs, SYL3C-modified MNPs, NC3S-modified MNPs, and the cocktail of aptamers and modified MNPs) (c). Comparison of capture efficiency of different aptamer modified MNPs for the mixture of MCF-7 and HeLa cells at a ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]1 (d).

To further evaluate the specificity performance of aptamer-modified MNPs for epithelial and mesenchymal CTCs, MCF-7 cells as a cell model of epithelial CTCs and HeLa cells as a cell model of mesenchymal CTCs were mixed at a ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]1 for the cell capture experiment. The results are shown in Fig. 4d. In the cell-mixture experiment, the capture efficiency of SYL3C-MNPs for MCF-7 and HeLa cells was 89.0% and 22.5%, respectively, whereas the NC3S-MNPs exhibited a different capture performance for MCF-7 cells (22.5%) and HeLa cells (91.3%). The results demonstrated that both NC3S and SYL3C aptamers possess an excellent capture specificity and efficiency for the target cells. Simultaneously, the capture yields of the aptamer–MNP cocktail (SYL3C-MNPs to NC3S-MNPs ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]1) were observed to respectively be 92.5% for MCF-7 cells and 92.0% for HeLa cells, indicating a fine capture performance of the aptamer–MNP cocktail for epithelial and mesenchymal CTCs.

Capture sensitivity of rare tumor cells in artificial CTC samples

We further investigated the capture sensitivity of NC3S-modified MNPs using a series of artificial CTC samples through spiking rare number HeLa cells (10, 20, 50, 100, and 200) into PBS or white blood cell suspension (WBCs) separated from 1 mL of fresh whole blood of healthy volunteers. The results are shown in Fig. 5a and b. More than 85% spiked HeLa cells could be isolated from the PBS samples and WBC mixture samples, indicating an excellent capture capacity of NC3S-modified MNPs toward the rare target cells.
image file: d0nr06180h-f5.tif
Fig. 5 The capture efficiency of NC3S-modified MNPs for rare number HeLa cells spiked into PBS (a) and WBC suspension (b). The capture efficiency of the aptamer–MNP cocktail for the mixed rare target cells (HeLa and MCF-7) in PBS (c) and WBC suspension (d).

To explore the potential clinical applications of the aptamer–MNP cocktail for rare target cell detection, 10, 20, 50, 100, and 200 of mixture cells were spiked into PBS or the WBC suspension. The mixture cells were composed of MCF-7 (prestained by DiI dye) and HeLa (prestained by DiO dye) cells at a constant ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]1. The results are shown in Fig. 5c and d. The aptamer–MNP cocktail exhibited a considerable capture efficiency and sensitivity for rare cancer cells of epithelial and mesenchymal phenotypes under different conditions.

Isolation of CTCs from blood samples of clinical patients

Finally, to evaluate the potential clinical utility of the aptamer–MNP cocktail for CTC isolation on peripheral blood samples from 6 healthy donors, 11 breast cancer patients, 1 rectal cancer patient, 2 colon cancer patients and 2 ovarian cancer patients (Table S1), the separated CTCs were identified by a typical immunofluorescence analysis using Hoechst 33342 (nuclear marker), Alexa Fluor 488-labeled anti-CD45 (CD45) (leukocyte marker), and Alexa Fluor 555-labeled anti-Pan-Keratin (PanCK) (epithelial marker) to distinguish CTCs among the contaminating WBCs. After the image acquisition, the cells showing Hoechst 33342+, CD45−, and PanCK+ were counted as CTCs, while the cells showing Hoechst 33342+, CD45+, and PanCK− were delineated as WBCs (Fig. 6a). We enumerated the CTCs isolated from these clinical blood samples, and the results are shown in Fig. 6b. No CTCs were found in blood samples from healthy donors, while for 15 out of 16 blood samples collected from breast cancer patients, rectal cancer patients, colon cancer patients and ovarian cancer patients, CTCs were successfully isolated by using the aptamer–MNP cocktail, indicating a hopeful clinical application of the aptamer–MNP cocktail for CTC isolation from clinical samples.
image file: d0nr06180h-f6.tif
Fig. 6 Isolation of CTCs by using the aptamer–MNP cocktail from blood samples of cancer patients. Immunofluorescence images of single CTC (Hoechst 33342+/CD45−/PanCK+) and WBC (Hoechst 33342+/CD45+/PanCK−) isolated from patient blood samples by using the aptamer–MNP cocktail, the scale bar in the images is 10 μm. (a). CTC count of blood samples obtained from 6 healthy donors (HD1-6), 11 breast cancer patients (BrC1-11), 1 rectal cancer patients (ReC1), 2 colon cancer patients (CoC1-2), and 2 ovarian cancer patients (OvC1-2) by the aptamer–MNP cocktail (b).

Conclusion

N-cadherin is highly expressed on the cell membrane of CTCs after EMT and can be used as a biomarker of CTCs with mesenchymal properties. However, CTC detection for the mesenchymal phenotype is mainly dependent on the anti-N-cadherin antibody, resulting in restricted applications. So it is urgent to develop an effective alternative for CTC detection. In this work, we successfully identified DNA aptamer targeting N-cadherin with high affinity and specificity after 12 rounds of selection. The aptamer NC3S showed the highest affinity with the lowest Kd value of 20.08 ± 0.72 nM. Flow cytometry and confocal imaging showed that the selected aptamers can specifically recognize a variety of cells expressing N-cadherin but do not bind to the negative cells. The two kinds of aptamer-modified magnetic nanoparticles (SYL3C-MNPs and NC3S-MNPs) that target two types of cancer cell surface biomarkers EpCAM and N-cadherin respectively showed an excellent capture specificity and efficiency for the target cells. And, the aptamer–MNP cocktail exhibited a considerable capture efficiency and sensitivity for rare cancer cells of epithelial and mesenchymal phenotypes. For the blood samples from patients, the CTCs were successfully isolated by using the aptamer–MNP cocktail. Given these, the dual-aptamer strategy could be used as a highly promising tool for the rapid detection and isolation of CTCs from clinical samples.

Conflicts of interest

The authors declare no competing financial interest.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (No. 21575154, 21904135, and 21775160), the CAS International Cooperation Key program (No. 121E32KYSB20170025) and the CAS/SAFEA International Innovation Teams program.

References

  1. C. Alix-Panabières and K. Pantel, Clin. Chem., 2013, 59, 110–118 CrossRef.
  2. K. A. Pramani, S. Jones, Y. Gao, C. Sweet, A. Vangara, S. Begum and P. C. Ray, Adv. Drug Delivery Rev., 2018, 125, 21–35 CrossRef.
  3. L. Cabel, C. Proudhon, H. Gortais, D. Loirat, F. Coussy, J. Y. Pierga and F. C. Bidard, Int. J. Clin. Oncol., 2017, 22, 421–430 CrossRef.
  4. C. M. Court, J. S. Ankeny, S. Sho, P. Winograd, S. Hou, M. Song, Z. A. Wainberg, M. D. Girgis, T. G. Graeber, V. G. Agopian, H. R. Tseng and J. S. Tomlinson, Ann. Surg. Oncol., 2018, 25, 1000–1008 CrossRef.
  5. P. S. Steeg and D. Theodorescu, Nat. Clin. Pract. Oncol., 2008, 5, 206–219 CrossRef CAS.
  6. B. Hong and Y. Zu, Theranostics, 2013, 3, 377–394 CrossRef CAS.
  7. J. J. Christiansen and A. K. Rajasekaran, Cancer Res., 2006, 66, 8319–8326 CrossRef CAS.
  8. X. Yan, L. Yan, S. Liu, Z. Shan, Y. Tian and Z. Jin, Mol. Med. Rep., 2015, 12, 2999–3006 CrossRef CAS.
  9. C. H. Lai, W. S. Tsai, M. H. Yang, T. Y. Chou and Y. C. Chang, Nanoscale, 2019, 11, 21119–21127 RSC.
  10. S. Wang, K. Liu, J. Liu, Z. T. Yu, X. Xu, L. Zhao, T. Lee, E. K. Lee, J. Reiss, Y. K. Lee, L. W. Chung, J. Huang, M. Rettig, D. Seligson, K. N. Duraiswamy, C. K. Shen and H. R. Tseng, Angew. Chem., Int. Ed., 2011, 50, 3084–3088 CrossRef CAS.
  11. S. Wang, H. Wang, J. Jiao, K. J. Chen, G. E. Owens, K. Kamei, J. Sun, D. J. Sherman, C. P. Behrenbruch, H. Wu and H. R. Tseng, Angew. Chem., Int. Ed., 2009, 48, 8970–8973 CrossRef CAS.
  12. L. Zhao, C. Tang, L. Xu, Z. Zhang, X. Li, H. Hu, S. Cheng, W. Zhou, M. Huang, A. Fong, B. Liu, H. R. Tseng, H. Gao, Y. Liu and X. Fang, Small, 2016, 12, 1072–1081 CrossRef CAS.
  13. J. I. Varillas, K. Chen, J. Zhang, T. J. George Jr. and Z. Fan, Methods Mol. Biol., 2017, 1634, 33–53 CrossRef CAS.
  14. S. Hao, Y. Wan, Y. Q. Xia, X. Zou and S. Zheng, Adv. Drug Delivery Rev., 2018, 125, 3–20 CrossRef CAS.
  15. Q. Huang, F. Wang, C. Yuan, Z. He, L. Rao, B. Cai, B. Chen, S. Jiang, Z. Li, J. Chen, W. Liu, F. Guo, Z. Ao, S. Chen and X. Zhao, Theranostics, 2018, 8, 1624–1635 CrossRef CAS.
  16. N. Sun, M. Liu, J. Wang, Z. Wang, X. Li, B. Jiang and R. Pei, Small, 2016, 12, 5090–5097 CrossRef CAS.
  17. A. Prodeus, A. Abdul-Wahid, N. W. Fischer, E. H. Huang, M. Cydzik and J. Gariepy, Mol. Ther. – Nucleic Acids, 2015, 4, e237 CrossRef CAS.
  18. Y. Han, D. Diao, Z. Lu, X. Li, Q. Guo, Y. Huo, Q. Xu, Y. Li, S. Cao, J. Wang, Y. Wang, J. Zhao, Z. Li, M. He, Z. Luo and X. Lou, Anal. Chem., 2017, 89, 5270–5277 CrossRef CAS.
  19. Y. Chen, H. Li, T. Gao, T. Zhang, L. Xu, B. Wang, J. Wang and R. Pei, Sens. Actuators, B, 2018, 254, 214–221 CrossRef CAS.
  20. T. Bing, N. Zhang and D. Shangguan, Adv. Biosyst., 2019, 3, e1900193 CrossRef.
  21. Z. Wang, N. Sun, M. Liu, Y. Cao, K. Wang, J. Wang and R. Pei, ACS Sens., 2017, 2, 547–552 CrossRef CAS.
  22. C. Tuerk and L. Gold, Science, 1990, 249, 505–510 CrossRef CAS.
  23. A. D. Ellington and J. W. Szostak, Nature, 1990, 346, 818–822 CrossRef CAS.
  24. Z. Zhuo, Y. Yu, M. Wang, J. Li, Z. Zhang, J. Liu, X. Wu, A. Lu, G. Zhang and B. Zhang, Int. J. Mol. Sci., 2017, 18, 2142–2160 CrossRef.
  25. S. Catuogno and C. L. Esposito, Biomedicines, 2017, 5, 49–66 CrossRef.
  26. P. Bayat, R. Nosrati, M. Alibolandi, H. Rafatpanah, K. Abnous, M. Khedri and M. Ramezani, Biochimie, 2018, 154, 132–155 CrossRef CAS.
  27. D. Shangguan, Y. Li, Z. Tang, Z. C. Cao, H. W. Chen, P. Mallikaratchy, K. Sefah, C. J. Yang and W. Tan, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 11838–11843 CrossRef CAS.
  28. Y. Song, Z. Zhu, Y. An, W. Zhang, H. Zhang, D. Liu, C. Yu, W. Duan and C. J. Yang, Anal. Chem., 2013, 85, 4141–4149 CrossRef CAS.
  29. J. W. Kim, E. Y. Kim, S. Y. Kim, S. K. Byun, D. Lee, K. J. Oh, W. K. Kim, B. S. Han, S. W. Chi, S. C. Lee and K. H. Bae, Mol. Cells, 2014, 37, 742–746 CrossRef.
  30. Z. Wang, N. Sun, H. Liu, C. Chen, P. Ding, X. Yue, H. Zou, C. Xing and R. Pei, ACS Appl. Mater. Interfaces, 2019, 11, 39586–39593 CrossRef CAS.
  31. A. Storck, J. Ludtke, L. Kopp and L. Juckem, BioTechniques, 2017, 63, 136–138 CrossRef CAS.
  32. D. Shangguan, L. Meng, Z. C. Cao, Z. Xiao, X. Fang, Y. Li, D. Cardona, R. P. Witek, C. Liu and W. Tan, Anal. Chem., 2008, 80, 721–728 CrossRef CAS.
  33. M. Takahashi, E. Sakota and Y. Nakamura, Biochimie, 2016, 131, 77–84 CrossRef CAS.
  34. M. Zuker, Nucleic Acids Res., 2003, 31, 3406–3415 CrossRef CAS.
  35. X. Guo, Z. Wu, W. Li, Z. Wang, Q. Li, F. Kong, H. Zhang, X. Zhu, Y. Du, Y. Jin, Y. Du and J. You, ACS Appl. Mater. Interfaces, 2016, 8, 3092–3106 CrossRef CAS.
  36. S. C. Gopinath, T. Lakshmipriya, Y. Chen, M. K. Arshad, J. P. Kerishnan, A. R. Ruslinda, Y. Al-Douri, C. H. Voon and U. Hashim, Appl. Microbiol. Biotechnol., 2016, 100, 6955–6969 CrossRef CAS.

Footnotes

Electronic supplementary information (ESI) available. See DOI: 10.1039/d0nr06180h
These two authors contributed equally.

This journal is © The Royal Society of Chemistry 2020