Construction and Assemble of miRNA Lockers

To develop a rapid and low-cost approach to generate miRNA Lockers, our strategy is to firstly assemble dsDNA devices containing the coding strand of miRNA Lockers using modularized DNA parts, then generate single strand miRNA Lockers by asymmetric PCR (Figure 1a). Nucleotides used for the assembly of dsDNA device can be found in our readily available nucleotide library. Our model can be used to assist selecting proper nucleotides for assembly.

For the rapid assembly of Locker-encoding dsDNAs, we modified the previously reported oligo-linker mediated assembly (OLMA) method to achieve high-throughput one-step assembly of multiple short dsDNA nucleotides. In our Overlapped Oligo Assembly (OOA) method, oligos with overlapped 5’-overhangs are synthesized, phosphorylated by nucleotide kinase and then ligated onto pSB1C3 plasmid backbone by T4 DNA ligase (Figure 1a). The correct assembling order is guaranteed by using identical overlap sequences among oligos (See Design Page).  As a primary validation, four experimental cases were constructed to examine the scalability, reliability and efficiency of the OOA method. Specifically, Locker A was designed with two identical functional modules to target miR-214, Locker B contained two different functional modules to target miR-214 and miR-654 at the same time, Locker C contained three identical modules to target miR-214, and Locker D contained four different modules targeting let-7a, miR-16, miR-195 and miR-184 (Figure 1b, left panel). Nucleotides chosen from the library encoding the supporting modules and functional modules of each of the four Locker designs were assembled using OAA method and transformed into E.coli DH5α chassis. Assemblies were validated via sequencing using universal primers for pSB1C3 backbone. Results showed >50% success rates in each case, with few mutations identified. (Figure 1b, middle panel).

Afterwards, to generate single strand miRNA Lockers, miRNA-Locker-coding region was digested from the plasmid and then used as the template for asymmetric PCR. To optimize the PCR protocol, primers in ratios of 20/1, 50/1 and 1/0 were tested. The asymmetric PCR products were analyzed through 3% agarose gel electrophoresis. As showed in the right panel of Figure 1b, bands showing ssDNA products were observed in all four cases when 20:1, 50:1 and 1:0 primer ratios were applied. Gray-scale analysis normalized with 20:1 group showed that the ssDNA amplification signal in 20:1 groups were generally 1.3-2-fold higher than 50:1 and 1:0 groups, suggesting that 20:1 might be an optimal primer ratio for ssDNA production. It is also notable that dsDNAs were also amplified even in 1:0 groups, which was thought to be the result of imperfect bindings of primers. Nevertheless, our results proved that ssDNA Lockers can be generated via OAA assembly followed with asymmetric PCR. Single strand DNAs can then be harvested and purified by standard PAGE purification assay.

Figure 1. Generation of miRNA Lockers with OAA assembly and asymmetric PCR. (a) Schematic representation of the Locker generating procedure. (b) Left panel: Design of four testing cases for the validation of the locker-generating procedure. Different miRNA binding modules were shown by different colors; middle panel: Sequencing validation for the ds DNA assemblies encoding four miRNA lockers; right panel: electrophoresis showing the production of four miRNA lockers by asymmetric PCR.

Validation of the miRNA inhibitory effect of miRNA Lockers

To verify the miRNA-inhibiting function of the miRNA Lockers, we tested our system with miR-214, a tumor-metastasis related miRNA. Previous reports have shown the repressive effect of miR-214 on the epithelial-mesenchymal transition (EMT) of cancer cells by targeting RNF8, an important EMT promoting protein (Figure 2a). Hence, the inhibitory effects on miR-214 can be observed through multiple manners including the abundance of miR-214 and its target gene, as well as RNF8 and EMT related phenotypes such as cell proliferation and migration. Here, we designed two types of miR-214 Locker as shown in Figure 2a. Locker-1(Lc-1) was designed containing two miR-214 binding sites, while Locker-2 (Lc-2) contained 3 binding sites. The control locker (Lc-C) was constructed by replacing the miRNA binding sites of Lc-1 with equal length of poly-C sequence to avoid unnecessary cross talks with other host miRNAs while maintaining the same structure.

To evaluate the miRNA binding ability of the Lockers in living cells, immunoprecipitation-PCR assay (IP-PCR) was performed to examine the RISC-mediated interaction between the Lockers and its target miRNA. Primers were designed on the supporting module amplifying one of the loop structure of the Locker to minimize the negative effect of highly-repeated miRNA binding sequences on the efficiency of PCR reaction (Figure 2b, upper scheme). For such assay, A549 cells transfected with specific sets of nucleotides were cultured for 36 hours and harvested for IP-PCR assay. As shown in Figure 2b, ~80bp (for Lc-1) or ~100bp (for Lc-2) DNA fragments was amplified from the precipitates of the groups co-transfecting Flag-tagged hAgo2 (Flag-hAgo2) expressing plasmid and Lc-1/Lc-2 by using anti-Flag antibody. By contrast, the negative control immunoprecipitations that did not use antibodies (Labeled as Beads) or adopted the normal rabbit IgG showed no amplification signal. Meanwhile, the control groups transfecting either miR-214 Lockers or Flag-hAgo2 expressing plasmid did not display any amplification signal. These results suggested that miRNA lockers were efficiently recognized by ago-miRNA complex in ex vivo cell cultures.

Then, to evaluate the effect of miR-214 Locker on miR-214 abundance in A549 cells, Taq-man based Real-Time Quantitative Reverse Transcription PCR (RT-qPCR) assay was performed. For such assay, A549 cells transfected with different Lockers or commercially purchased microRNA inhibitor were cultured for 36 hours before harvesting. Locker Lc-C was used as the negative control for Lockers, while commercially purchased antagomir for miR214 (An-214) was used as the positive control. The Control antagomir molecule (An-C) was used as the control group for An-214. Significant reduction of miR-214 abundance was observed in both Lc-1 and Lc-2 groups comparing to the Lc-C group with ~80%-90% reduction rates (Figure 2c). Similar phenomenon was also observed on antagomir groups with a lower reduction rate (around 60%, Figure 2c), which was in accordance with previous report. Such results suggested a better inhibitory effect of miR-214 Lockers comparing to commercially purchased product.

Subsequently, Western blot analysis was performed to examine the effect of miR214 Lockers on the expression level of RNF8, a previously reported gene regulated by miR-214. Results showed that RNF8 levels in Lc-1 and Lc-2 group increased for 1.32 and 1.49 folds comparing to the Lc-C group, while the RNF8 expression level in An-214 group increased for 1.17 folds comparing to the control group (Figure 2d). Such results demonstrated the capability of miR214 Lockers on affecting the expression level of RNF8, a miR-214 regulated gene, with a similar but more effective manner comparing to miRNA antagomirs.

Since miRNA inhibitor, such as Antagomirs, was normally utilized to miRNA loss of function researches. For Further verification, we examined whether miRNA Locker generate miRNA knock-down-corelated phenotype changes identical to those generated by commercially purchased miRNA inhibitors.

Knock-down of miR-214 expression level has been demonstrated to promote the growth of lung adenocarcinoma cells and be involved in migration and invasion of lung cancer via downregulation of RNF8. We then tested if the miR-214 Lockers can induce phenotype changes as expected. For such matter, cell proliferation assay using Cell Counting Kit-8 (CCK8) was performed to evaluate the effect of miR-214 Locker on cell proliferation. As showed in Figure 2f, the absorbance of 450nm (indicating the living cell number) in Lc-2 transfected A549 cells were significantly increased comparing to the Lc-C groups, indicating that Lc-2 transfection could significantly strengthen cell proliferation capacity (Figure 2e). Furthermore, Transwell analysis was performed to examine the effect of miR-214 Lockers on the migration of A549 cells. Results showed significant increases of cell numbers migrated through Transwell chamber in Lc-1 and Lc-2 groups comparing the Lc-C group (Figure 2f), while miR-214 antagomir (An-214) showed similar effect. Cell counts demonstrated a ~2-fold increase in Lc-1 group and a ~3-fold increase in Lc-2 group comparing to Lc-C group, while a ~2-fold increase is also observed in the An-214 group comparing to the control group (Figure 2g). These results suggested the capability of miR-214 Locker to induce phenotype changes in a miR-214-antagomir-similar pattern.

Figure. 2 Validation of the miRNA inhibitory effect of miRNA Lockers (a) Schematic representation showing the effect of miR-214 on EMT process of cancer cells. Two miRNA Lockers, Lc-1 and Lc-2 are expected to promote EMT by blocking the repressive effect of miR-214 on RNF8, A EMT promoting protein. (b) IP-PCR assay determining the binding ability of miRNA Lockers with target microRNA in A549 cells. Schematic representation of primer design is shown above, hAgo2 stands for the Flag-tagged human Ago2 expressing plasmid, input indicates an aliquot of total DNA. Antibodies used for immunoprecipitation are indicated above the lanes. (c) RT-qPCR results demonstrating miR-214 abundance in A549 cells transfected with miR-214 Lockers Lc-C (as control) /Lc-1/Lc-2 or miR-214 Antagomir (An-214)/Antagomir Control (An-C). (d)Western blot determining RNF8 expression level in A549 cells transfected with miR-214 Lockers Lc-C/Lc-1/Lc-2 or Antagomir An-214/An-C. (e) CCK8 Assay indicating the proliferation of A549 cells transfected with miR-214 Lockers Lc-C/Lc-1/Lc-2. (f)and (g) Tranwell assay determining of migration in A549 cells transfected with miR-214 Lockers Lc-C/Lc-1/Lc-2 or Antagomir An-214/An-C. Relative gene expression was calculated using the 2^-ΔΔCT method, with initial normalization of genes against U6 snRNA within each group. The expression levels of each gene in the control groups (Lc-C or An-C) were arbitrarily set to 1.0. Relative protein expression levels were calculated by using β-actin expression level as initial normalization and then set the protein expression level in the control groups (Lc-C or An-C) arbitrarily as 1.0. The relative expression values were averaged from the data in three parallel reactions, and the results were obtained from at least three independent experiments. Error bars represent SD.

Utilizing miRNA Locker to determine the effect of miR-654 in the EMT regulation of human lung adenocarcinoma cells.

Here, we demonstrated how miRNA Lockers can be used in miRNA researches as a promising substitute of current miRNA inhibitors by identifying a new RNF8-targeting miRNA and establish its regulatory relationship with EMT.

By setting RNF8 as the target downstream gene in three different computational algorithms that analyze miRNA binding sites on the 3’-UTR of a specific gene (TargetScan, miRBase and miRWalk), miR-654-5p (miR-654 for short) was predicted as a promising candidate regulator of RNF8 by directly bind with its 3’-UTR (Data not shown). However, such regulatory relationship has not been experimentally validated yet, while the relationship between miR-654 and cancer metastasis also remains poorly discussed. Hence, we designed four Lockers for miR-654, named as Lc-1p，Lc-1np，Lc-2p and Lc-2np. Locker-1p (Lc-1p) and Locker-1np (Lc-1np) contained two miR-654 binding sites, while Locker-2p (Lc-2p) and Lc-2np (Lc-2np) contained three miR-654 binding sites. Suffix ‘p’ and ‘np’ was marked to distinguish the non-perfect complementary base pairing (np) and perfect complementary base pairing (p) between miRNA Locker and target miRNA (figure 3a). The control locker (Lc-C) was constructed similarly as the control locker of miR-214 by replacing the miRNA binding sites of Lc-1p with equal length of poly-C sequence to avoid unnecessary cross talks with other host miRNAs while maintain the same structure.

As such, IP-PCR was performed to verify the miR-654 binding ability of the designed Lockers. Similarly, primers were designed on the supporting module amplifying one of the loop structure of the Lockers. A549 cells transfected with specific sets of nucleotides were cultured for 36 hours and harvested for IP-PCR assay. As shown in Figure 3b, ~80bp (for Lc-1p and Lc-1np) or ~100bp (for Lc-2p and Lc-2np) DNA fragments was amplified from the precipitates of the groups co-transfected with Flag-hAgo2 expressing plasmid and Lc-1p/Lc-1np/Lc-2p/Lc-2np miRNA Lockers by using anti-Flag antibody, while the negative control that did not use antibodies or adopted the normal rabbit IgG showed no amplification signal. Meanwhile, the control group transfected with Flag-hAgo2 expressing plasmid alone did not display any amplification signal. Besides, the knockdown effect of miR-654 Lockers on miR-654 in A549 cells were validated through Taqman-based RT-qPCR assay. Results showed that miR-654 level significantly reduced in four miR-654 Lockers’ groups comparing to the Lc-C group with ~80%-98% reduction rates (Figure 3c), indicating a good inhibitory effect of miR-654 Lockers on miR-654. Similar phenomenon could be observed on antagomir groups with a lower reduction rate (around 60%, Figure 3c). Our results suggested that our fine designed miRNA lockers were efficiently recognized by ago-miRNA complex and further functioned by lowering the target miRNA level.

Since prediction results suggested that miR-654 might downregulate RNF8 expression by directly targeting 3’UTR of RNF8 mRNA. To validate such prediction, Western blot analysis was performed to examine the protein expression level of RNF8 under miR-654 knock-down condition. As showed in Figure 3d, western Blot results showed 2-fold to 3.5-fold increases on RNF8 protein level in Locker-supplemented groups comparing to the Lc-C group, while the RNF8 expression level in An-654 group only increased for 1.17 folds comparing to the control group. In addition, miR-654 knockdown reduced epithelial marker (E-cadhein) and increased mesenchymal marker (Snail, N-cadhein) (Figure 3e), which suggested that Locker-mediated miR-654 knockdown might be related with enhanced EMT via upregulation of RNF8. Taking together, these results demonstrated miRNA Locker-induced knockdown of miR-654 could increase expression of RNF8 and plausibly promote EMT.

With previous results, we could conclude that miR-654 participant in the regulation of RNF8 expression in A549 cells. Since it has been proved that RNF8 participated in cell proliferation and migration in non-small lung cancer cell A549. We anticipate miR-654 might be involved in migration and proliferation of lung cancer by regulating RNF8 expression. To further study the role of miR-654 in A549 proliferation and migration, Cell proliferation assay using Cell Counting Kit-8 (CCK8) was performed to detect the migration capacity of A549 cells transfected with four types of miR-654 Lockers or with Lc-C control. Results showed that the absorbance of 450nm (indicating the living cell number) in four Lockers-groups were significantly increased comparing to the Lc-C groups (figure 3f), indicating that the knockdown of miR-654 could significantly strengthen cell proliferation capacity. Subsequently, Transwell assay was carried out to test the migration capacity of Locker-mediated miR-654 knockdown groups. Compared with control, miR-654 knock-down dramatically increased the number of migrated A549 cells with 3.5~5-fold changes (Figure 3h), suggesting that knockdown of miR-654 accelerates cell migration of non-small cell lung cancer A549 cell. Transwell results of miR-654 knockdown by antagomir (~2-fold change) confirmed such conclusion. Taken together, we found that block of miR-654 using our miRNA lcoker promoted cell proliferation and migration of lung adenocarcinoma cells.

Figure 3. Utilizing miRNA Locker to determine the effect of miR-654 in the EMT regulation of human lung adenocarcinoma cells.  (a) Schematic representation showing the anticipated effect of miR-654 on EMT process of human lung adenocarcinoma cells, the relation between miR-654 and RNF8 is unverified. Four miRNA Lockers, Lc-1p, Lc-1np, Lc-2p, Lc-2np are expected to promote EMT by blocking the possible repressive effect of miR-654 on RNF8.  (b) IP-PCR assay determining the binding ability of miRNA Lockers with miRNA in A549 cells. Schematic representation of primer design is shown above. Each group was co-transfected corresponding Lockers/Lc-C with hAgo2 expression plasmid. Input indicates an aliquot of total DNA. Antibodies used for immunoprecipitation are indicated above the lanes. (c) RT-qPCR results demonstrating miR-654 abundance in A549 cells transfected with miR-654 Lockers Lc-C/Lc-1p/Lc-1np/Lc-2p/Lc-2np or Antagomir An-654/An-C. (d) RT-qPCR results showing RNF8 mRNA expression level in A549 cells transfected with miR-654 Lockers Lc-C/Lc-1p/Lc-1np/Lc-2p/Lc-2np or Antagomir An-654/An-C. (e) Western blot analysis determining the expression level of epithelial marker E-cadherin and mesenchymal markers (N-cadherin and Snail) in A549 cells transfected with miR-654 Lockers Lc-C/Lc-1p/Lc-1np/Lc-2p/Lc-2np or Antagomir An-654/An-C. (f) CCK8 Assay indicating the proliferation of A549 cells transfected with Lockers Lc-C/ Lc-1p/Lc-1np/Lc-2p/Lc-2np (g) and (h) Tranwell assay determining of migration in A549 cells transfected with miR-654 Lockers Lc-C/Lc-1p/Lc-1np/Lc-2p/Lc-2np or Antagomir An-654/An-C. Relative gene expression was calculated using the 2^-ΔΔCT method, with initial normalization of genes against U6 snRNA within each group. The expression levels of each gene in the control groups (Lc-C or An-C) were arbitrarily set to 1.0. Relative protein expression levels were calculated by using β-actin expression level as initial normalization and then set the protein expression level in the control groups (Lc-C or An-C) arbitrarily as 1.0. The relative expression values were averaged from the data in three parallel reactions, and the results were obtained from at least three independent experiments. Error bars represent SD.

Discussion and Future work

In summary, our results suggested that the miR-214 Lockers were capable of binding and inhibiting miR-214 in ex vivo lung cancer cell cultures effectively. We also demonstrated that miR-214 Locker could induce downstream gene expression changes and phenotype changes in the similar pattern as existing miRNA antagomir. Interestingly, generally better inhibitory effects were observed in miR-214 Lockers comparing to commercially purchased miRNA inhibitor. We further demonstrated how miRNA Lockers can be used in miRNA researches as a promising substitute of current miRNA inhibitors by identifying a new RNF8-targeting miRNA and establish its regulatory relationship with EMT. These results indicated that miRNA Lockers designed following our principle could be a promising substitute for existing miRNA inhibitor with similar or even stronger effects comparing to existing ones while significantly reduce the costs.

Validation of the miRNA inhibitory effect of miRNA Lockers

(1)To optimized the conditions of assembly reaction and improve efficiency of assembly.

(2)To assemble multi-miRNAs Locker with different miRNA binding sites.

(3)To test cytotoxicity of assembled miRNA Locker

(4)To evaluatethe possibility and safety of using miRNA Locker in vivo.  

(5)Toexplore the detailed mechanism of miRNA Locker inhibiting miRNA level.