A general approach to high-yield biosynthesis of chimeric RNAs bearing various types of functional small RNAs for broad applications

Abstract

RNA research and therapy relies primarily on synthetic RNAs. We employed recombinant RNA technology toward large-scale production of pre-miRNA agents in bacteria, but found the majority of target RNAs were not or negligibly expressed. We thus developed a novel strategy to achieve consistent high-yield biosynthesis of chimeric RNAs carrying various small RNAs (e.g. miRNAs, siRNAs and RNA aptamers), which was based upon an optimal noncoding RNA scaffold (OnRS) derived from tRNA fusion pre-miR-34a (tRNA/mir-34a). Multi-milligrams of chimeric RNAs (e.g. OnRS/miR-124, OnRS/GFP-siRNA, OnRS/Neg (scrambled RNA) and OnRS/MGA (malachite green aptamer)) were readily obtained from 1 l bacterial culture. Deep sequencing analyses revealed that mature miR-124 and target GFP-siRNA were selectively released from chimeric RNAs in human cells. Consequently, OnRS/miR-124 was active in suppressing miR-124 target gene expression and controlling cellular processes, and OnRS/GFP-siRNA was effective in knocking down GFP mRNA levels and fluorescent intensity in ES-2/GFP cells and GFP-transgenic mice. Furthermore, the OnRS/MGA sensor offered a specific strong fluorescence upon binding MG, which was utilized as label-free substrate to accurately determine serum RNase activities in pancreatic cancer patients. These results demonstrate that OnRS-based bioengineering is a common, robust and versatile strategy to assemble various types of small RNAs for broad applications.

Figure 1.

High-yield production of recombinant miRNA/siRNA agents in E. coli using OnRS-based technology. (a) Urea–PAGE analysis of total RNAs showed that there was large variability in the expression of chimeric pre-miRNAs in E. coli using the same tRNA scaffold. Total RNAs isolated from untransformed HST08 (WT) E. coli were used as a control. The heat color gradation indicates the base-pairing probability from 0 to 1. (b) The chimeric tRNA/mir-34a robustly expressed in E. coli was developed as an OnRS that offered a consistent high-level expression of chimeric miRNAs (e.g. OnRS/miR-124) and siRNAs (e.g. OnRS/GFP-siRNA) in E. coli (e.g. ∼15–20% of total RNAs). In contrast, there was no or minimal expression of miR-124 and GFP siRNA using the tRNA and tRNA/mir-155 scaffold, respectively. (c and d) Representative FPLC traces during the purification of OnRS/miR-124 and OnRS/GFP-siRNA, respectively. Inserts are corresponding urea–PAGE analyses of collected fractions (1, 2, 3 and 4) eluted at 8.3 and 8.7 min, respectively, which confirmed the purity of isolated recombinant ncRNAs.

Figure 2.

Fate of recombinant ncRNAs in human cells. (a and b) Unbiased deep sequencing study revealed that OnRS-carried miR-124 and GFP-siRNA were precisely processed to target small RNAs, leading to 3 orders of magnitude increase in miR-124 in A549 cells and GFP siRNA in ES-2/GFP cells, respectively. Note the presence of miRNA and siRNA isoforms as well as corresponding passenger strands and other small RNAs at much lower levels (Supplementary Tables S2 and S4). In contrast, the levels of other cellular miRNAs showed no or minor changes (Supplementary Tables S3 and S5). Values are mean ± SD of triplicated treatments that were sequenced separately. (c and d) Mapping major cellular tRFs derived from OnRS/miR-124 versus OnRS (tRNA/mir-34a) in A549 cells and OnRS/GFP-siRNA versus OnRS/Neg in ES-2/GFP cells, respectively. Shown are the mean numbers of reads of triplicated treatments. 3000 and 1000 reads was used as a cut off for the A549 and ES-2/GFP cells, respectively. See the Supplementary Tables S2 and S4 for a complete list of tRFs derived from each ncRNA.

Figure 3.

OnRS-carried miRNA is biologically/pharmacologically active in regulating target gene expression and controlling cellular processes in human cells. (a) RT-qPCR analysis revealed that mature miR-124 levels retained 3 orders of magnitude higher in A549 cells for 4 days since transfection with OnRS/miR-124, as compared with OnRS/Neg. (b) Western blots showed that OnRS/miR-124 was effective in reducing the protein expression level of miR-124 target gene STAT3 in A549 cells at 72 h post-transfection. (c) Flow cytometric analyses demonstrated that OnRS/miR-124 was effective in inducing apoptosis in A549 cells at 48 h post-transfection. Cells treated with OnRS/Neg were used as controls. (d) MTT assay showed that OnRS/miR-124 significantly suppressed the proliferation of A549 cells at 72 h post-treatment, as compared to OnRS/Neg. (e) Inhibition of A549 cell proliferation by OnRS/miR-124 was also demonstrated when cell growth was monitored using Icelligence Real-Time Cell analyzer. The arrow points to the time point of ncRNA treatment. Values are mean ± SD of triplicated treatments. *P < 0.01.

Figure 4.

OnRS-carried siRNA is effective for RNAi in vitro and in vivo. GFP fluorescence intensity was sharply reduced in ES-2/GFP cells in vitro at 72 h after transfected with OnRS/GFP-siRNA (a), which was associated with (b) 70–80% lower GFP mRNA levels and (c) 1000-fold higher GFP siRNA levels. Following i.v. administration of OnRS/GFP-siRNA, hepatic GFP fluorescence was significantly suppressed in the GFP-transgenic mouse models in vivo, as demonstrated by microscopic examination of (d) non-fixed and (e) fixed liver slices, as well as (f) RT-qPCR analysis of hepatic GFP mRNA levels. Fixed liver slices were stained with DAPI, and GFP fluorescence and DAPI-stained nuclei (blue) images were merged together (e). Control ES-2/GFP cells (N = 3 per group) or GFP-transgenic mice (N = 3–4 per group) were treated with the same doses of OnRS/Neg. Values are mean ± SD. *P < 0.01, compared with OnRS/Neg treatment.

Figure 5.

High-yield production of functional RNA aptamers in bacteria using OnRS. (a) Representative design of OnRS/aptamer forms where RNA aptamer was inserted at the 5′ or 3′ end of hsa-mir-34a. The heat color gradation indicates the base-pairing probability from 0 to 1. (b) A consistent high-level expression of OnRS-carried MGA in E. coli, i.e. over 50% of OnRS/MGA5 and OnRS/MGA3 in total RNAs. (c) Representative FPLC traces of OnRS/MGA5 during FPLC purification. Insert is urea–PAGE analysis of the collected RNA fractions (1, 2 and 3) eluted at 10.6 min. (d) Binding to OnRS/MGA5 and OnRS/MGA3 led to a shift of the wavelength of MG maximum absorbance from 618 to 630 nm. The same shift was observed when FPLC-purified OnRS/MGA and total RNAs isolated from OnRS/MGA-expressing bacteria were used. The sephadex aptamer (OnRS/Seph) and corresponding total RNAs were used as additional controls. (e) Strong and selective fluorescence was shown when MG bound to OnRS/MGA5 or OnRS/MGA3. The same results were obtained when using FPLC-purified OnRS/MGA and OnRS/MGA-containing total RNAs.

Figure 6.

Application of label-free, OnRS-carried MGA sensor to the determination of serum RNase activities in pancreatic cancer patients. (a) Change in MGA-bound-MG fluorescent intensity with the increase in MG and MGA concentrations. Corresponding MGA and MG concentrations were fixed at 1.6 μg/ml and 10 μM, respectively. (b) The fluorescent intensity was decreased over time when incubated with human serum, and in vivo-jetPEI formulated OnRS/MGA was protected from cleavage by serum RNases. (c) Dose response was obvious for the exposure to human serum RNases and the intensity of OnRS/MGA-bound MG fluorescence, and addition of RNase inhibitor completely blocked the cleavage of OnRS/MGA by serum RNases. (d) OnRS/MGA was much more susceptible to human RNase A (10 min incubation) than angiogenin (RNase 5; 30 min incubation). (e) Human pancreatic cancer patients showed significantly higher serum RNase activities than benign/normal patients, as determined by the decrease in MGA-bound MG fluorescence intensity (△A.U./min/μl). N = 10 in each group. OnRS/MGA5 was used in this study.