Building a chimera of aptamer–antisense oligonucleotide for silencing galectin-1 gene

Abstract

Galectin-1 is closely related with immune systems, and its overexpression may cause tumor metastasis. Owing to the better stability of antisense oligonucleotides than siRNA, we constructed a chimera of aptamer–antisense oligonucleotide (Apt–AS) for silencing the galectin-1 gene specifically. For studying the endocytosis pathways of Apt–AS by aptamer targeted delivery, we marked the 3′ terminus of Apt–AS with Alexa Flour 488 (Apt–AS488) so as to trace the pathways through the confocal microscope. Furthermore, Apt–AS was used to silence the expression of galectin-1. The results showed Apt–AS entered cells by a caveolae mediated endocytosis pathway, and the aptamer–antisense chimeras did not impair the antisense gene silencing efficiency. Moreover Apt–AS could improve the cellular uptake and selectively entered cells.

---

Introduction

As DNA nanotechnology evolves, DNA nanostructures are increasingly being used in targeted drug delivery as smart nanomaterials.^1–3 Most DNA nanostructures can be made to assemble themselves with intermolecular forces such as hydrogen bonding and π-stacking,^4,5 and some structures can also interact with some specific proteins or molecules.^6,7 These DNA structures have good biocompatibility, and some can easily penetrate into cells. Aptamer is one of the typical representatives.⁸ All of these particular merits make the aptamers wide use in tumor detection, therapy, bioimaging, biosensor and so on.^7,9

Galectin is the member of lectin, which can target β-galactoside sugars specifically and share some amino acid sequence motif.¹⁰ Galectin-1 is one of the types of galectin family, and plays an important role in nervous system and the stability of T-cells in vivo.¹¹ Moreover, the overexpression of galectin-1 in tumors or the tissues surrounding them should be supposed to a sign of more advanced diseases.¹² These are also connected with the cancer metastasis, which may spread to the surrounding healthy tissues or escape from the immune system.¹²

Winkler and co workers¹³ have reported that the DNA antisense oligonucleotide (TTCGTATCCATCTGGCAGC) could silence the gene of galectin-1, by entering into cells with the help of the Lipofectamine 2000. The general strategy to silence the functional genes is to use siRNA or DNA antisense oligonucleotides to target the messenger RNA of functional genes. The delivery of these oligonucleotides into cells are mainly by carriers and aptamers.^14,15 The carriers often include virus vectors, polyethylenimine (PEI), cationic lipid and other nanoparticles. Recently, Chang and co-workers¹⁶ used the third generation of lentiviral vector to co-transfect the plasmid DNA and siRNA to silence astrocyte elevated gene-1. Aigner and co-workers¹⁷ had constructed a PEI–siRNA complex to silence vascular endothelial growth factor gene. Winkler used Lipofectamine 2000 (the most common used cationic lipid) to deliver the siRNA and miRNA-like oligonucleotides.¹³ These carriers can enter into cells efficiently, but lacking specificity and biocompatibility.

Aptamers are obtained from random nucleotides sequence pool through SELEX,¹⁸ which can self-assemble to second structures and possess high specificity and affinity to bind target molecules.^19–21 In 2006, the first aptamer–siRNA chimera was constructed, and the authors used the complete oligonucleotide-based method to deliver siRNA, in which the RNA aptamer could target PSMA membrane protein to enter tumor cell and the chimeras would be processed by Dicer, then depleting the siRNA targeting protein.²² The aptamer portion of aptamer–siRNA chimeras has the promise to selectively drive the siRNA portion to the cancer tissues or cells that overexpress the aptamer targeting proteins, and the siRNA potion binds with the targeted messenger RNA to lead to the specific cleavage of mRNA and efficient silencing of the targeted gene expression.^23,24 With the development of aptamers, the DNA aptamer–siRNA chimeras were also applied because of the higher chemical stability and lower possibility to trigger the non-specific effect mediated by interferon.²⁵ For examples, AS1411 aptamer–luciferase SSO chimeras could target nucleolin and bind the targeted pre-mRNA.²⁶ The multivalent comb-type aptamer–siRNA conjugates (MUC1 aptamer–Bcl-2 siRNA chimeras) were delivered into cells and silenced Bcl-2 gene which resulted in massive apoptosis.²⁷ DNA aptamer–siRNA chimera was reported to enter into CD4+ T cells specifically and then silence the expression of HIV-PR gene.²⁵

Antisense and siRNA oligonucleotides are quite similar in physicochemical properties, but they differ in the molecular machinery.¹³ Rudin and co-workers²⁸ used liposome to delivery c-raf-1 antisense oligonucleotides. Other delivery systems were also reported. Tan and co-workers⁹ constructed a multifunctional DNA nanomaterial AptNA by base pair hybridization, in which the aptamer domain could drive AptNA into CEM cells and silence P-glycoprotein gene.

However, to our knowledge, so far no aptamer–DNA antisense oligonucleotide chimera has been reported. The advantage of chimera structures is that they consist of pure nucleotides except adding a chemical spacer, which could ensure the firm conjugation between aptamers and siRNA so that the two portions could work in cooperation. Furthermore, oligonucleotides chimeras could be directly synthesized on nucleic acids synthesis machine, which makes the process of preparing them more straightforward and versatile.²⁹ And the synthesis process and the stability of DNA antisense oligonucleotides are superior to siRNA.

From these researches, we got inspired to construct a DNA nanostructure Apt–AS (Scheme 1), which contains aptamer domain, spacer, and antisense domain. The aptamer domain is AS1411, an intermolecular G-quadruplex,³⁰ which can target nucleolin specifically. Nucleolin is overexpressed on membrane of the some tumor cells so that it can be an ideal tumor-specific target for delivery. But in the contrast, nucleolin is distributed at nucleus in normal cells.³¹ The spacer is used to separate aptamer from antisense domain to keep activity of aptamer and antisense oligonucleotide. And the antisense domain is for gene silencing of galectin-1. In order to investigate the endocytosis pathways of Apt–AS, we marked Alexa flour488 at the 3′ terminus (Apt–AS488). We also performed the gene silencing assays to evaluate the effect of Apt–AS to the expression of galectin-1.

Scheme 1 Schematic of the G-quadruplex formation of Apt–AS.

Experimental

Materials

DNA oligonucleotides were obtained from Sangon Biotechnology Co., Ltd. (Shanghai, China). All of the DNA oligonucleotides were prepared in ultrapure water, and then used the 260 nm UV absorbance to accurately quantify the concentrations. Polyethylenimine (PEI) 25KD, methyl-β-cyclodextrin (MβCD), sucrose, chlorpromazine hydrochloride (CPZ) were purchased from Sigma-Aldrich. Cytochalasin D (Cyt D) was purchased from Aladdin (Shanghai, China). Lyso-tracker red, Dual Luciferase Reporter Gene Assay Kit and Cell Counting Kit-8 (CCK-8) were purchased from Beyotime Institute of Biotechnology. DMEM, Opti-MEM® I medium and fetal bovine serum (FBS) were purchased from Gibco. The plasmid psiCHECK2 was obtained from Promega, and the Homo sapiens galectin-1 cDNA was purchased from Sino Biological Inc (Beijing, China). XhoI and NotI were purchased from Thermo scientific Oligo dT primers (PrimeScript™ RT Reagent Kit) and SYBR Green preMix (TakaRa SYBR Premix EX Taq Kit) were purchased from TaKaRa Biological Inc (Dalian, China).

Cell culture and image analysis

Breast cancer cell line (MCF-7), metastatic lung cancer cell line (A549) and human embryonic kidney cell line (293T) were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and cultured in DMEM with 10% FBS and 1% penicillin/streptomycin. All the cells were cultured in 37 °C and 5% CO₂. For cell imaging, cells were seeded in 35 mm glass-bottomed culture dishes one day in advance. When the cells grew up to 60–70% confluency, the culture mediums was replaced with a certain amount of Apt–AS488 or AS488 in Opti-MEM® I medium. Then the culture dishes were incubated in 37 °C for 2 hours. Before imaging, cells were washed with PBS for three times and added DMEM (10% FBS plus). Immediately, the cells were visualized in confocal laser scanning microscopy (TCS SP5, Leica, Germany). Objective: 63×.

Endocytosis pathway analysis

Cells were pretreated with 10 μM CytD, 10 mM MβCD, 450 mM sucrose or 10 μg mL⁻¹ chlorpromazine hydrochloride at 37 °C for 30 min, respectively. After washed with PBS, Apt–AS488 in Opti-MEM® I medium was added into glass-bottomed culture dishes for 2 hours at 37 °C. For energy dependent endocytosis inhibition assay, cells were pretreated in 4 °C for 30 min, and DNA uptake processes were also performed by incubation in 4 °C. The results were analyzed in BD Accuri C6 cytometer (USA) and visualized in confocal laser scanning microscopy.

Flow cytometry

Cells were seeded in 24-well plate at density of 7 × 10⁴ cells per well one day in advance. When the cells grew up to 80% confluency, the culture mediums was replaced with a certain amount of Apt–AS488 or AS488 in Opti-MEM® I medium. Then the cells were incubated in 37 °C for 2 hours. Later, the cells were washed three times with PBS and were added 200 μL trypsin for 1 min at 37 °C. Next, 500 μL DMEM (10% FBS plus) was added to suspend cells and transferred to 1.5 mL tubes. BD Accuri C6 cytometer was used to analyze the fluorescence of cells, and at least 10 000 cells were collected.

Construction of recombination plasmid psiCHECK2-gal-1

According to the report by Nasrin Shokrzadeh et al., we constructed a recombination plasmid psiCHECK2-gal-1, which was established by cloning the Homo sapiens galectin-1 cDNA into psiCHECK2 plasmid vector.¹³ The Homo sapiens galectin-1 cDNA, was used as template for PCR-amplifying the galectin-1 cDNA fragment with restrict enzyme sites of XhoI and NotI. Then, the galectin-1 cDNA fragment and the vector plasmid psiCHECK2 were digested with XhoI and NotI, respectively. And the plasmid psiCHECK2-gal-1 was established by ligating with the sticky ends of the digested galectin-1 cDNA and psiCHECK2 fragments. The resulting plasmid was sequenced by Genewiz to confirm the correct sequence of psiCHECK2-gal-1.

Gene silencing assay

The psiCHECK2-gal-1 has two reporter genes. The firefly luciferase intensity was used as a control for normalization, and the renilla luciferase evaluated the interaction between antisense and target gene. For a positive control, antisense and psiCHECK2-gal-1 plasmid were co-transfected by PEI. The Dual Luciferase Assay Kit was used to measure the luminescence intensities of firefly luciferase and renilla luciferase after 48 h. For the experimental group, the psiCHECK2-gal-1 plasmid was firstly transfected by PEI for 24 h, and then the medium was replaced by fresh DMEM medium with 10% FBS containing amounts of antisense (Apt–AS or AS). Likewise, the Dual Luciferase Assay Kit was used to measure the luminescence intensities after 48 h by Perkin Elmer Victor™ X4 Multilabel Plate Reader (USA). The normalized relative luminescence was calculated by the following equation:

The normalized relative luminescence (%) = (R_sample − R_blank)/(R_control − R_blank)

R_sample represents the ratio between the luminescence intensities of renilla luciferase and firefly luciferase when being added Apt–AS or AS.

R_blank represents the ratio between the luminescence intensities of renilla luciferase and firefly luciferase without cells and any treatment.

R_control represents the ratio between the luminescence intensities of renilla luciferase and firefly luciferase when cells were only treated with plasmid.

Cell proliferation assay

The cell proliferation of Apt–AS against MCF-7 cells was evaluated by Cell Counting kit-8 (CCK-8). The MCF-7 cells were seeded in 96-well plate at density of 7000 cells per well. When cells grew up to 60–70%, the culture medium was replaced with different concentrations of Apt–AS in DMEM (10% FBS plus) and incubated in 37 °C for 24 hours. Then, 10 μL of CCK-8 was added to each well and incubated for further 2 h. The absorbance was measured at 450 nm using a microplate reader (Perkin Elmer Victor X4). The cell proliferation rates were calculated according to the following equation:

Cell proliferation rate (%) = 100 × (OD_sample − OD_background)/(OD_control − OD_background)

RNA extraction and qPCR analysis

The total RNA of cells from the positive control and experimental group was isolated by TRIzol method (Invitrogen, Carlsbad, CA, USA). The concentration of total RNA was measured by 260 nm UV absorbance, and 500 ng of total RNA was transcribed into cDNA using Oligo dT primers. Applied Biosystem 7500 was used to perform the quantitative polymerase chain reaction (qPCR) with SYBR Green preMix. The fold change or relative expression normalized to β-actin endogenous control was used and the SYBR green primers used for qPCR assay are listed in ESI Table S1.† The relative quantification (untreated cells to treated cells) was determined by following equation:

ΔCt_control = Ct_control − Ct_actin,

ΔCt_sample = Ct_sample − Ct_actin,

ΔΔCt = ΔCt_control − ΔCt_sample

Relative change in mRNA levels = 2^−ΔΔCt

Results and discussions

The selectivity of aptamer–antisense chimera to different cell lines

AS1411 aptamer could target nucleolin specifically, which over-expressed in the membrane of some tumor cells. And as such, we compared the target efficiency of Apt–AS488 when incubating with different cells. MCF-7 and A549 are typical tumor cells that overexpress nucleolin in the cell membrane. Fig. 1a showed the fluorescence intensity of MCF-7 was higher than A549 and 293T, and 293T was the lowest. So we hypothesized that the distribution of nucleolin on MCF-7 is more than on A549 cells. In addition, 293T cells belong to normal cells so that the nucleolin localizes in nucleus. Fig. 1b and 2 showed the fluorescence intensity of Apt–AS488 was higher than AS488 in the same cell line, and Apt–AS488 should enter cells more readily than the antisense without aptamer (AS488). Therefore, Apt–AS488 has the characteristics of high specificity and selectivity.

Fig. 1 (a) The relative increasing fluorescence intensity of MCF-7, A549 and 293T cells after incubation with Apt–AS488. (b) The relative increasing fluorescence intensity of MCF-7 cells after incubation with Apt–AS488 and AS488. Data were obtained by flow cytometry analysis and subtracted the fluorescence of blank cells. The concentration of Apt–AS488 and AS488 is 400 nM, and the concentration of Apt–AS and AS is 400 nM (*p < 0.1, **p < 0.05, ***p < 0.001, n = 6).

Fig. 2 Confocal fluorescence images of MCF-7 cells after incubation with AS488 (b) and Apt–AS488 (c) for 2 h, respectively. (a) is the blank control. Excitation wavelength was set at 488 nm, and the emission wavelength is 510–580 nm, scale bar: 20 μm. The concentration of Apt–AS488 and AS488 is 400 nM.

Endocytosis pathway analysis

After incubating with Apt–AS488, we stained cells with Lyso-tracker red for co-localization with lysosomes. As shown in Fig. 3, we could find some Apt–AS488 dispersing in lysosomes. If Apt–AS488 entered the cells through energy dependent endocytosis pathway, we should consider the receptor mediated endocytosis. And the receptor mediated endocytosis pathways could be divided into three types: clathrin mediated endocytosis, caveolae mediated endocytosis and the above independent receptor mediated endocytosis. Firstly, we study the energy dependent endocytosis by temperature experiments. Cells were cultured in 4 °C for 30 min, and then incubated with Apt–AS488 for 2 h in 4 °C. All of energy dependent endocytosis pathways were terminated when cells were incubated in 4 °C. From Fig. 4 and S1a–d,† the fluorescence intensity of cells at 4 °C decreased apparently, compared with cells at 37 °C. The inhibition ratio of 4 °C was 90.84% ± 1.53%. This demonstrated Apt–AS488 entered into cells by energy dependent endocytosis. Furthermore, the cytochalasin D (Cyt D) was used to investigate the non-receptor mediated endocytosis. The results showed that the uptake of Apt–AS488 had no significant decline when the cells were pretreated with Cyt D and then incubated with Apt–AS488. Thus, we assumed that the Apt–AS488 entered cells by receptor mediated endocytosis pathway.

Fig. 3 Confocal fluorescence images corresponding to co-localization of Apt–AS488 (green channel) and Lyso-tracker red (red channel) in MCF-7 cells after incubation with Apt–AS488 for 2 h and then with Lyso-tracker red for 15 min (b). And (a) is the blank control. Overlay image demonstrated that Apt–AS488 located both in lysosome and cytoplasm, scale bar: 20 μm. The concentration of Apt–AS488 is 400 nM.

Fig. 4 Endocytosis pathway analysis of Apt–AS488 when entering into MCF-7 cells. (a) The fluorescence intensity of MCF-7 cells after incubation with 10 mM MβCD, 10 μM Cyt D, 450 mM sucrose, and 10 μg mL⁻¹ CPZ before internalization of Apt–AS488. 37 °C means MCF-7 cells were incubation with Apt–AS488 at 37 °C without any pretreation. And 4 °C means MCF-7 cells were pretreated in 4 °C for 30 min, and then incubated with Apt–AS488 in 4 °C. (b) Corresponding inhibition ratios. The concentration of Apt–AS488 is 400 nM (*p < 0.1, ***p < 0.001, n = 6).

The clathrin mediated endocytosis and caveolae mediated endocytosis were the primary types of receptor mediated endocytosis pathway.³² In order to study the role of these two pathways in uptaking Apt–AS488 into cells, we inhibited one of the two pathways by the inhibitor. As an inhibitor of caveolae mediated endocytosis, MβCD could reduce cholesterol and destroyed the formation of caveolae.^33,34 Cells were pretreated with MβCD, and then incubated with Apt–AS488. The results of confocal imaging and flow cytometry constantly showed MβCD could inhibit the uptake of Apt–AS488, and the internalized ratio was 36.86% ± 4.25% (Fig. 4, S1e and f†). As for clathrin mediated endocytosis pathway, we used sucrose and chlorpromazine (CPZ) to be the inhibitors. Cells were pretreated with these two inhibitors, and then incubated with Apt–AS488. However, these two inhibition assays both showed that the amount of Apt–AS488 entering cells had no obvious decrease (Fig. S1i–l†). Thus, we could conclude that Apt–AS488 entered cells by caveolae mediated endocytosis pathway.

Gene silencing activity assay

In order to silence gene galectin-1, we constructed a recombination plasmid psiCHECK2-gal-1, which was fused with the human cDNA of galectin-1. The firefly luciferase intensity was used as control for normalization, and the renilla luciferase evaluated the interaction between antisense and target gene.¹³ For positive control, Apt–AS and AS were respectively mixed with psiCHECK2-gal-1. The mixture was co-transfected into MCF-7 by PEI, the Dual Luciferase Assay Kit from Promega was used to measure the luminescence intensities after 48 h. From Fig. 5a, the gene silencing efficiency of Apt–AS was basically in line with AS, which showed that adding aptamer did not impair the gene silencing efficiency of antisense, and Apt–AS could exert the interference function normally. For the application of oligonucleotides without transfection reagent, the psiCHECK2-gal-1 plasmid was firstly transfected by PEI, then the mixture solution was removed after 24 h, and the fresh medium containing the Apt–AS or AS was replaced. Luminescence intensities were measured after 48 h incubation. Fig. 5b showed both the gene silencing efficiency of Apt–AS and AS was decreased comparing with the positive control, but the normalized relative luminescence of Apt–AS was significantly lower than AS. We also performed qPCR to confirm the expression level of galectin-1 mRNA (Fig. S2†), and the results were consistent with the dual luciferase reporter assay. Apt–AS could selectively target the nucleolin on the MCF-7 cells surface, and then it was uptaked by cells through the caveolae mediated endocytosis pathway. Apt–AS would escape during the maturity of lysosome, so that it could silence the galectin-1 gene in psiCHECK2-gal-1 plasmid. Thus, Apt–AS could selectively target the nucleolin on the MCF-7 cells surface to improve the cellular uptake of oligonucleotides without the transfection reagents.

Fig. 5 Effect of Apt–AS and AS in a dual luciferase reporter assay after transfection with PEI (a) and after naked application (b) (*p < 0.1, **p < 0.05, n = 6).

Cell proliferation assay

For excluding the effect of anti-proliferative activity of AS1411, cell proliferation was evaluated after treatment with the different concentrations of Apt–AS (Fig. 6). From the results, the cell proliferation rates of MCF-7 did not decrease significantly with increasing the concentration of Apt–AS. The cell proliferation rate of Apt–AS against MCF-7 cells was nearly 100% even at the Apt–AS concentration of 2 μM. This result demonstrated that Apt–AS possesses well biocompatibility.

Fig. 6 CCK-8 assays of Apt–AS in different concentrations against MCF-7 cells for 24 h (*p < 0.1, **p < 0.05, ***p < 0.001, n = 6).

Conclusions

Aptamer–antisense oligonucleotide chimera was firstly constructed in gene silencing assay. For tracking the cellular uptake process of Apt–AS, we labelled Alexa flour 488 (Apt–AS488) in the 3′ terminus. We found Apt–AS488 could enter into cells selectively. We also studied the endocytosis pathways of Apt–AS488 into the mammalian cells, and the result showed Apt–AS488 could enter the cells by caveolae mediated endocytosis pathway. For the application of Apt–AS, the gene silencing assay was carried out. When Apt–AS or AS was co-transfected with psiCHECKE2-gal-1 by PEI, the gene silencing progress was happening along with the endocytosis. We found Apt–AS did not impair the antisense gene silencing efficiency. As for the naked application of Apt–AS, psiCHECKE2-gal-1 was firstly transfected for stabilization in the cells, then incubated with Apt–AS or AS. It turned out that Apt–AS could improve the cellular uptake and entered cell selectively. The aptamer domain showed important advantages over transfection reagents for antisense delivery. Although most of the commercial gene deliverable regents like PEI have high efficacy, they do not have the ability to target cells selectively. And the use of Apt–AS provides the possibility to overcome these problems enabling the accumulation of antisense oligonucleotides specifically in MCF-7 cells in a safe and effective manner, so that Apt–AS could play an important role in gene silencing.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (21575154, 21507156) and the CAS/SAFEA International Innovation Teams program.

Notes and references

1. A. H. Okholm and J. Kjems, Adv. Drug Delivery Rev., 2016, 106, 183–191.
2. N. C. Seeman, Nature, 2013, 421, 427–431.
3. V. Linko, A. Ora and M. A. Kostiainen, Trends Biotechnol., 2015, 33, 586–594.
4. J. M. Heemstra, Beilstein J. Org. Chem., 2015, 11, 2713–2720.
5. R. Chhabra, J. Sharma, Y. Liu, S. Rinker and H. Yan, Adv. Drug Delivery Rev., 2010, 62, 617–625.
6. A. Czogalla, H. G. Franquelim and P. Schwille, Biophys. J., 2016, 110, 1698–1707.
7. V. Kumar, S. Palazzolo, S. Bayda, G. Corona, G. Toffoli and F. Rizzolio, Theranostics, 2016, 6, 710–725.
8. S. Cansiz, L. Zhang, C. Wu, Y. Wu, I. T. Teng, W. Hou, Y. Wang, S. Wan, R. Cai, C. Jin, Q. Liu and W. Tan, Chem.–Asian J., 2015, 10, 2084–2094.
9. C. Wu, D. Han, T. Chen, L. Peng, G. Zhu, M. You, L. Qiu, K. Sefah, X. Zhang and W. Tan, J. Am. Chem. Soc., 2013, 135, 18644–18650.
10. S. H. Barondes, D. N. Cooper, M. A. Gitt and H. Leffler, J. Biol. Chem., 1994, 269, 20807–20810.
11. H. Blanchard, K. Bum-Erdene, M. H. Bohari and X. Yu, Expert Opin. Ther. Pat., 2016, 26, 537–554.
12. I. Camby, M. Le Mercier, F. Lefranc and R. Kiss, Glycobiology, 2006, 16, 137R–157R.
13. N. Shokrzadeh, A. M. Winkler, M. Dirin and J. Winkler, Bioorg. Med. Chem. Lett., 2014, 24, 5758–5761.
14. Y. Diao, J. Liu, Y. Ma, M. Su, H. Zhang and X. Hao, Cancer Biol. Ther., 2016, 17, 498–506.
15. J. Zhou and J. J. Rossi, Silence, 2010, 1, 4.
16. Y. Chang, B. Li, X. Xu, L. Shen, H. Bai, F. Gao, Z. Zhang and J. B. Jonas, PLoS One, 2016, 11, e0148763.
17. S. Hobel, I. Koburger, M. John, F. Czubayko, P. Hadwiger, H. P. Vornlocher and A. Aigner, J. Gene Med., 2010, 12, 287–300.
18. A. D. Ellington and J. W. Szostak, Nature, 1990, 346, 818–822.
19. Y. X. Wu and Y. J. Kwon, Methods, 2016, 106, 21–28.
20. 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.
21. H. M. Meng, H. Liu, H. Kuai, R. Peng, L. Mo and X. B. Zhang, Chem. Soc. Rev., 2016, 45, 2583–2602.
22. J. O. McNamara 2nd, E. R. Andrechek, Y. Wang, K. D. Viles, R. E. Rempel, E. Gilboa, B. A. Sullenger and P. H. Giangrande, Nat. Biotechnol., 2006, 24, 1005–1015.
23. C. L. Esposito, S. Catuogno and V. de Franciscis, J. RNAi Gene Silencing, 2014, 10, 500–506.
24. H. Y. Liu, X. Yu, H. Liu, D. Wu and J. X. She, Sci. Rep., 2016, 6, 30346.
25. Q. Zhu, T. Shibata, T. Kabashima and M. Kai, Eur. J. Med. Chem., 2012, 56, 396–399.
26. J. W. Kotula, E. D. Pratico, X. Ming, O. Nakagawa, R. L. Juliano and B. A. Sullenger, Nucleic Acid Ther., 2012, 22, 187–195.
27. H. Yoo, H. Jung, S. A. Kim and H. Mok, Chem. Commun., 2014, 50, 6765–6767.
28. C. M. Rudin, J. L. Marshall, C. H. Huang, H. L. Kindler, C. Zhang, D. Kumar, P. C. Gokhale, J. Steinberg, S. Wanaski, U. N. Kasid and M. J. Ratain, Clin. Cancer Res., 2004, 10, 7244–7251.
29. R. L. Juliano, X. Ming, K. Carver and B. Laing, Nucleic Acid Ther., 2014, 24, 101–113.
30. L. Xu, X. Shen, S. Hong, J. Wang, Y. Zhang, H. Wang, J. Zhang and R. Pei, Chem. Commun., 2015, 51, 8165–8168.
31. A. S. Barbas, J. Mi, B. M. Clary and R. R. White, Future Oncol., 2010, 6, 1117–1126.
32. V. Haucke, Nature, 2015, 517, 446–447.
33. L. Liang, J. Li, Q. Li, Q. Huang, J. Shi, H. Yan and C. Fan, Angew. Chem., 2014, 53, 7745–7750.
34. M. Yang, X. Zhang, H. Liu, H. Kang, Z. Zhu, W. Yang and W. Tan, Anal. Chem., 2015, 87, 5854–5859.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details and other data. See DOI: 10.1039/c6ra21250f

---