Communication pubs.acs.org/JACS
Genetically Encoded Fluorescent RNA Sensor for Ratiometric Imaging of MicroRNA in Living Tumor Cells Zhan-Ming Ying, Zhan Wu, Bin Tu, Weihong Tan, and Jian-Hui Jiang* Institute of Chemical Biology and Nanomedicine, State Key Laboratory of Chemo/Bio-Sensing and Chemometrics, College of Chemistry and Chemical Engineering, Hunan University, Changsha 410082, People’s Republic of China S Supporting Information *
also been developed for high-contrast imaging applications in bacteria toward varying targets including metabolites,13,14 proteins15 and RNA.16 However, light-up aptamers have rarely been realized for RNA imaging in living mammalian cells, possibly because of relatively low abundance and poor stability of these RNA aptamers in mammalian cells.17 Moreover, lightup RNA sensors have only allowed imaging using a single-color fluorescence, indicating that current methods still suffer from increased variability in detection of cellular targets.17 Novel designs of light-up RNA sensors that enable ratiometric imaging are in high demand for quantitative live cell studies. To address these issues, we report the development of a novel genetically encoded RNA sensor for fluorescent imaging of microRNAs (miRs) in living mammalian (tumor) cells. By coexpression of the RNA sensor with GFP, we also realize this RNA sensor for dual-emission ratiometric imaging in living cells. MiRs are known as major players in post-transcriptional gene regulation, and their aberrant expression is closely associated with many diseases such as cancers.18 Nevertheless, light-up RNA sensors for ratiometric imaging of miRs in living cells have been largely unexplored. We reasoned that RNA aptamers with stable secondary structure could exhibit resistance to misfolding and show desirable stability in tumor cells, affording the potential for engineering a genetically encoded RNA sensor for miRs. Motivated by this hypothesis, we chose the well-developed RNA aptamer for fluorophore sulforhodamine B (SR) to design the miR-activated genetically encoded sensor. The SR-binding aptamer was reported to have a three-way junction structure and light up quenched fluorescence of a conjugate of SR with dinitroaniline (DN).19 This property was demonstrated for fluorescence imaging of RNA molecules fused with the SR aptamer.19 However, it has not been developed as a light-up sensor for cellular imaging. To construct an SR-binding aptamer-based sensor for a miR target, a responsive motif needed to be incorporated in a folding-dependent domain of the aptamer. To this end, we utilized the basic principle of molecular beacon operation2−5 and developed a novel genetically encoded light-up RNA sensor using a stem−loopshaped miR-responsive motif, as illustrated in Scheme 1A. A prototype sensor was designed by trimming the SR-binding aptamer in the terminal stem, followed by incorporation with a miR-responsive stem−loop motif. This design destroyed the structure for the SR-binding loop, which prevented the
ABSTRACT: Light-up RNA aptamers are valuable tools for fluorescence imaging of RNA in living cells and thus for elucidating RNA functions and dynamics. However, no light-up RNA sensor has been reported for imaging of microRNAs (miRs) in mammalian cells. We report a novel genetically encoded RNA sensor for fluorescent imaging of miRs in living tumor cells using a light-up RNA aptamer that binds to sulforhodamine and separates it from a conjugated contact quencher. On the basis of the structural switching mechanism for molecular beacon, we show that the RNA sensor activates high-contrast fluorescence from the sulforhodamine-quencher conjugate when its stem− loop responsive motif hybridizes with target miR. The RNA sensor can be stably expressed within a designed tRNA scaffold in tumor cells and deliver light-up response to miR target. We also realize the RNA sensor for dualemission, ratiometric imaging by coexpression of RNA sensor with GFP, enabling quantitative studies of target miR in living cells. Our design may provide a new paradigm for developing robust, sensitive light-up RNA sensors for RNA imaging applications.
F
luorescence imaging of RNA in living cells is essential for understanding the functions and dynamics of diverse classes of cellular RNAs encoded by the genome.1 Current methods mostly rely on the use of green fluorescent protein (GFP)-tagged RNA binding proteins, or hybridization probes such as molecular beacon2−5 and spheric nucleic acids (SNA)based nanoflares.6,7 Despite their success for visualizing the expression and transport of RNA in living cells, these methods still suffer from limitations such as interferences from probes on functions of RNA, difficulty in efficiently delivering probes to target sites, or high background fluorescence from excess GFP and degraded probes.8 Development of low-background, robust and simple methods for live-cell RNA imaging remains a central challenge. Light-up RNA aptamers,9 especially recent discovery of the “spinach” system,10,11 have created a simple but valuable tool for live-cell imaging of RNAs. The light-up RNA aptamers are a class of engineered RNA aptamers able to turn on intense fluorescence upon binding to nonfluorescent, small-molecule dyes, thereby permitted direct tracking of RNA with low background and high sensitivity via the incorporated dyebinding motifs.12 By fusing the aptamer with target-responsive RNA modules, genetically encoded “signal on” sensors have © 2017 American Chemical Society
Received: May 3, 2017 Published: July 17, 2017 9779
DOI: 10.1021/jacs.7b04527 J. Am. Chem. Soc. 2017, 139, 9779−9782
Communication
Journal of the American Chemical Society
expression as well as the perturbations of operation factors such as excitation intensity. Compared to a recent Spinach aptamer-based RNA sensor that used two separate, aptamerinterspaced sequences for miR hybridization, and only afforded compromised sensitivity with micromolar detection ranges,22 our design may be especially better suited for miR detection. MiRs only have ∼22 nucleotides, a sequence too short to form an energetically favorable duplex with two separate sequences, which could lead to failures in restoring the dye-binding loop. Our molecular beacon-based design with a single continuum loop for miR hybridization exhibits higher energetic stability, thereby conferring better sensitivity and signal-to-background ratio for miR detection. Therefore, our design may provide a new paradigm for developing robust, sensitive and convenient light-up sensors for imaging of miRs in living mammalian cells. We chose miR-21, a known biomarker overexpressed in many cancers,23 as a case of study to demonstrate our method. The SR−DN conjugate was synthesized according to a reported procedure and characterized using NMR and MS (Figure S1).19 The designed RNA sequences were prepared by in vitro RNA transcription using T7 RNA polymerase and the corresponding DNA templates (Scheme S1). A preliminary optimization experiment involving optimization of the trimmed SR-binding aptamers revealed that a four base-pair stem was adequate for stabilization of the dye-binding loop (Figure S2). On the basis of the trimmed SR-binding aptamer, RNA sensors were constructed, and in vitro assays for the as-prepared RNA sensors were examined (Figure 1A). The tRNA-supported
Scheme 1. Illustration of Genetically Encoded RNA Sensor for Ratiometric miR Imaginga
a
(A) Design of light-up RNA sensor. (B) Realization of genetically encoded sensor in tumor cells.
prototype RNA sensor from lighting up the SR−DN conjugate. To achieve efficient expression of the RNA sensor in tumor cells, we engineered the aptamer sensor in a recombinant tRNA.20 The tRNA structure could act as a protective threedimensional scaffold, which improved resistance of the engineered sensor to misfolding and nuclease-mediated degradation.21 In the absence of target miR, the tRNAsupported RNA sensor could not bind to the SR−DN conjugate with a quenched fluorescence. When target miR hybridized with its complementary loop in the miR-responsive motif, the masked stem sequence was released, allowing spontaneous folding of the SR-binding loop. As a result, binding of the properly folded aptamer to the SR−DN conjugate activated an enhanced fluorescence signal for the detection of miR. To construct a light-up RNA sensor for miR imaging in living tumor cells, we further engineered it as a genetically encoded sensor with coexpression of GFP for dual-emission ratiometric detection, as illustrated in Scheme 1B. A plasmid for the sensor was constructed with two separate promoters: one to express the tRNA-supported RNA sensor and the other to express GFP. After transfection in the cells, the plasmid generated coexpression of the tRNA-supported RNA sensor with GFP. The RNA sensor delivers an orange fluorescence signal in response to target miR, whereas GFP gives target-independent green fluorescence as an internal reference. This dual-emission property permits ratiometric imaging for target miR in the cells. Unlike the design of plasimid coexpressing the Spinach aptamer with fluorescence protein for measurement of mRNA expression at the transcription and translation levels,21 our design for coexpression of RNA sensors with GFP creates a new approach enabling ratiometic imaging using light-up RNA aptamers. Ratiometric imaging mitigates experimental variability by normalizing the efficiency of transfection and
Figure 1. (A) Fluorescence spectra for SR−DN conjugate, RNA sensor and RNA sensor plus miR-21. (B) Fluorescence intensities of RNA sensor at 598 nm to other cellular components. (C) Fluorescence spectra of RNA sensor to miR-21 of varying concentration. (D) Gel electrophoresis image for RNA sensor with or without miR-21 stained using SR−DN conjugate or SYBR Green II. M denotes molecular weight markers.
sensor, obtained by incorporating the prototype sensor in the anticodon stem region of tRNAlys3,20 was observed to only exhibit slightly increased fluorescence on interacting with the SR−DN conjugate, suggesting collapse of the SR-binding loop. In the presence of miR-21 (5 μM), the fluorescence peak showed further ∼8-fold enhancement, an indicator of refolding of the SR-binding aptamer. Moreover, the tRNA-supported 9780
DOI: 10.1021/jacs.7b04527 J. Am. Chem. Soc. 2017, 139, 9779−9782
Communication
Journal of the American Chemical Society sensor did not substantially alter the responsive performance of the prototype sensor (Figure S3), verifying successful construction of the tRNA-supported sensor for miR-21. Further investigation of the tRNA-supported sensor in response to varying cellular components, such as some other miRs, total tRNA, proteins and lysate of negative cell L02,24 showed high selectivity to miR-21 (Figure 1B). In addition, the light-up sensor displayed enhanced fluorescence responses to miR-21 in a dose dependent manner in the concentration range of 1 nM to 5 μM (Figure 1C and Figure S4). The detection limit was estimated to be 0.3 nM, indicating high sensitivity of the sensor for miR-21 detection. Gel electrophoresis analysis revealed that the tRNA-supported RNA sensor showed decreased mobility upon interacting with miR-21 (Figure 1D and Figure S5). Furthermore, only the low-mobility band delivered bright fluorescence after staining with the conjugate. This result gave direct evidence for stable hybridization of the sensor with miR21 and selective fluorescence activation of the sensor by the target. Next, we applied the developed sensor for imaging of miR-21 in living tumor cells. To confirm the permeability of the SR− DN conjugate through cellular membrane, we incubated HeLa cells using the conjugate. Fluorescence imaging with a high concentration of SR−DN (20 μM) revealed that SR−DN was able to permeate through the membrane and display a predominant localization in the cytosol (Figure S6), verifying the feasibility of engineering light-up RNA sensors using the SR−DN conjugate for live-cell imaging. Moreover, at an incubation concentration of 1 μM SR−DN, negligible fluorescence was observed in the cells, which indicated that a working concentration of 1 μM SR−DN afforded very low fluorescence background and thus enabled high-contrast imaging using the light-up RNA sensors. Further interrogations revealed that bright fluorescence images were obtained in miR21 overexpressed HeLa cells24 after transfection using a plasmid composed of a U6 promoter and a downstream sequence encoding the tRNA-supported RNA aptamer or the tRNAsupported sensor (Figure S7). Moreover, the sensor delivered increasing fluorescence contrast with increased concentration of miR-21 in the cells (Figure S8). This result confirmed that the RNA aptamer and the sensor were stably expressed in HeLa cells, and that the sensor retained its responsiveness to miR-21 in the cells. Although tRNA-supported RNAs were also used for the Spinach system,10,11 it is noteworthy that no intense fluorescence signal was obtained in mammalian cells.17 A consistent result was also obtained in our initial experiments for the RNA sensor using Spinach aptamer, in which the sensor exhibited activated fluorescence in vitro but failed to show fluorescence signals in living tumor cells. This result might be ascribed to misfolding of the Spinach aptamer in tumor cells. Interestingly, when the RNA aptamer was not incorporated in the tRNA scaffold, much weaker fluorescence images were obtained (Figure S9). This finding gave clear evidence for the role of tRNA scaffolds in protecting the incorporated RNA sequences from misfolding or degradation. A reverse transcription polymerase chain reaction (RT-PCR) analysis showed that expression of the tRNA-supported sensor in the cells maintained a stable expression after transfection for 24 through 48 h (Figure S10). Therefore, we used cells collected at 24 to 48 h after transfection for subsequent studies. On the basis of the preliminary results, we constructed a plasmid with two promoters for dual-emission ratiometric imaging of miR-21 in tumor cells (Figure 2). The pRNAT-U6.1
Figure 2. Fluorescence images for HeLa cells (A) treated with 300 nM miR-21 inhibitor (B) and 300 nM miR-21 mimic (C).
plasmid with two promoters, U6 and CMV, was used for expression of the RNA sensor and GFP, respectively (Scheme S2). After HeLa cells were transfected with the plasmid, bright and uniform green fluorescence images appeared in the nuclei and in the cytosol, although the intensities were slightly different in these two areas. Such a uniform distribution of GFP in the nuclei or the cytosol afforded an ideal internal standard signal for ratiometric imaging. Moreover, bright red fluorescence (pseudocolor for orange channel) was observed in the cytosol in the HeLa cells, which were known to have overexpression of miR-21.24 This result indicated that the RNA sensor delivered activated fluorescence response to miR21. A control experiment using a control plasmid with the RNA sensor replaced by the tRNA scaffold reveal a very low background, confirming there was little nonspecific staining of cellular components by SR−DN (Figure S11 in SI). To further validate the specificity of the fluorescence response, a control experiment was performed using HeLa cells treated with miR21 inhibitor, small chemically modified single-stranded RNAs designed to specifically bind to and selective decrease active concentration of miR-21.25 As anticipated, cells treated with the inhibitor displayed very dim red fluorescence in the cytosol, implying the specificity of the RNA sensor in response to miR21. Another control experiment with HeLa cells treated with miR-21 mimic, small chemically modified double-stranded RNAs mimicking miR-21 and enabling up-regulation of miR-21 activity,25 showed much more intense red fluorescence in the cytosol than that obtained with the untreated cells. These findings revealed that the red fluorescence signal increased with increasing miR-21 concentration. The ratiometric images also exhibited a color shift depending on the concentration of miR21. We also observed cells with no fluorescence signals in the merged images, which were ascribed to the limited transfection efficiency. Because limited transfection efficiency unavoidably led to nontransfected cells with no expression of GFP and RNA sensors, this result actually suggested that our dual-emission design allowed directly assessing the transfection efficiency and mitigating the false negative responses. Closer interrogations of HeLa cells with varying miR-21 concentrations demonstrated ratiometric images with ratios of intensity (ROIs) that were dynamically correlated with the concentrations of miR-21, implying that the ratiometric images enabled quantitative studies of miR-21 in living cells (Figure S12). A further investigation of HeLa cells after transfection for 24 to 48 h showed that the red and green fluorescence signals remained almost unchanged, suggesting the stability of our RNA sensor for live-cell imaging of miR-21 (Figure S13). Additionally, a z-axis scan study clearly indicated that the red 9781
DOI: 10.1021/jacs.7b04527 J. Am. Chem. Soc. 2017, 139, 9779−9782
Communication
Journal of the American Chemical Society and the green fluorescence signals were both localized in the cells (Figure S14). Next, the potential of the RNA sensor for quantitative evaluation of miR-21 in different cell lines, MCF-7, Hela and L02, was examined (Figure 3A). We observed fluorescence
■
construction, RT-qPCR and fluorescence imaging as well as additional figures (PDF)
AUTHOR INFORMATION
Corresponding Author
*
[email protected] ORCID
Weihong Tan: 0000-0002-8066-1524 Jian-Hui Jiang: 0000-0003-1594-4023 Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS This work was supported by NSFC (21527810, 21521063).
Figure 3. (A) Fluorescence imaging for different cells. (B) RT-PCR analysis for the cells. (C) Calculated relative expression with reference to U6 and ROIs for the cells.
images of varying brightness and ratiometric images of different colors in these cells, and MCF-7 cells gave the highest miR-21 expression whereas L-02 had the lowest expression. This result was consistent with previous study for miR-21 expression in these cells.23 Moreover, the ROIs for red-to-green fluorescence were dynamically correlated with the relative expression levels of miR-21, as determined by RT-PCR (Figure 3B,C). These results verified the ability of our RNA sensor for quantitative imaging of miRs in living cells. In conclusion, we reported a novel genetically encoded RNA sensor for fluorescent imaging of miRs in living tumor cells based on the well-folded light-up RNA aptamer for an SR−DN conjugate. We demonstrated that engineering the light-up RNA aptamer with a stem−loop miR-responsive motif could enable development of a RNA sensor for selective and sensitive detection of miRs in different tumor cells. We also showed that insertion of this well-folded light-up RNA aptamer and the RNA sensor in a tRNA scaffold conferred adequate stability and abundance in tumor cells for fluorescence imaging. Moreover, the RNA sensor was realized for dual-emission ratiometric imaging by coexpression of the RNA sensor with GFP, and the obtained ratiometric images exhibited dynamic correlation with target miR concentrations. Although this genetically encoded RNA sensor still suffers from the effect of limited transfection efficiency, our design may open the possibility for developing robust, sensitive genetically encoded RNA sensors for visualizing miRs and elucidating their functions in living tumor cells.
■
REFERENCES
(1) Dean, K. M.; Palmer, A. E. Nat. Chem. Biol. 2014, 10, 512. (2) Zheng, J.; Yang, R.; Shi, M.; Wu, C.; Fang, X.; Li, Y.; Li, J.; Tan, W. Chem. Soc. Rev. 2015, 44, 3036. (3) Liu, J.; Cao, Z.; Lu, Y. Chem. Rev. 2009, 109, 1948. (4) Zhang, H.; Li, F.; Dever, B.; Li, X. F.; Le, X. C. Chem. Rev. 2013, 113, 2812. (5) Nutiu, R.; Li, Y. J. Am. Chem. Soc. 2003, 125, 4771. (6) Cutler, J. I.; Auyeung, E.; Mirkin, C. A. J. Am. Chem. Soc. 2012, 134, 1376. (7) Halo, T. L.; McMahon, K. M.; Angeloni, N. L.; Xu, Y.; Wang, W.; Chinen, A. B.; Malin, D.; Strekalova, E.; Cryns, V.; Cheng, C. H.; Thaxton, C. S.; Mirkin, C. A. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 17104. (8) Tyagi, S. Nat. Methods 2009, 6, 331. (9) Babendure, J. R.; Adams, S. R.; Tsien, R. Y. J. Am. Chem. Soc. 2003, 125, 14716. (10) Paige, J. S.; Wu, K. Y.; Jaffrey, S. R. Science 2011, 333, 642. (11) Strack, R. L.; Disney, M. D.; Jaffrey, S. R. Nat. Methods 2013, 10, 1219. (12) Hofer, K.; Langejürgen, L. V.; Jaschke, A. J. Am. Chem. Soc. 2013, 135, 13692. (13) Kellenberger, C. A.; Wilson, S. C.; Sales-Lee, J.; Hammond, M. C. J. Am. Chem. Soc. 2013, 135, 4906. (14) Paige, J. S.; Nguyen-Duc, T.; Song, W.; Jaffrey, S. R. Science 2012, 335, 1194. (15) Su, Y.; Hickey, S. F.; Keyser, S. G.; Hammond, M. C. J. Am. Chem. Soc. 2016, 138, 7040. (16) Sato, S. I.; Watanabe, M.; Katsuda, Y.; Murata, A.; Wang, D. O.; Uesugi, M. Angew. Chem., Int. Ed. 2015, 54, 1855. (17) Strack, R. L.; Song, W.; Jaffrey, S. R. Nat. Protoc. 2014, 9, 146. (18) Dong, H.; Lei, J.; Ding, L.; Wen, Y.; Ju, H.; Zhang, X. Chem. Rev. 2013, 113, 6207. (19) Sunbul, M.; Jäschke, A. Angew. Chem., Int. Ed. 2013, 52, 13401. (20) Ponchon, L.; Dardel, F. Nat. Methods 2007, 4, 571. (21) Pothoulakis, G.; Ceroni, F.; Reeve, B.; Ellis, T. ACS Synth. Biol. 2014, 3, 182. (22) Note: During the revision of the paper, we noticed a justpublished paper on RNA sensor for miR detection: Huang, Q.; Doyle, F.; Wurz, Z. E.; Tenenbaum, S. A.; Hammond, R. K.; Caplan, J. L.; Meyers, B. C. Nucleic Acids Res. 2017, DOI: 10.1093/nar/gkx504. (23) Lu, J.; Getz, G.; Miska, E. A.; Alvarez-Saavedra, E.; Lamb, J.; Peck, D.; Sweet-Cordero, A.; Ebert, B. L.; Mak, R. H.; Ferrando, A. A.; et al. Nature 2005, 435, 834. (24) Chan, J. A.; Krichevsky, A. M.; Kosik, K. S. Cancer Res. 2005, 65, 6029. (25) Leone, E.; Morelli, E.; Di Martino, M. T.; Amodio, N.; Foresta, U.; Gullà, A.; Rossi, M.; Neri, A.; Giordano, A.; Munshi, N. C.; et al. Clin. Cancer Res. 2013, 19, 2096.
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b04527. Experimental methods including in vitro transcription, fluorescence and gel electrophoresis analysis, plasmid 9782
DOI: 10.1021/jacs.7b04527 J. Am. Chem. Soc. 2017, 139, 9779−9782
