Labeling RNAs in live cells using malachite green aptamer scaffolds as fluorescent probes

RNAs mediate many different processes that are central to cellular function. The ability to quantify or image RNAs in live cells is very useful in elucidating such functions of RNA. RNA aptamerfluorogen systems have been increasingly used in labeling RNAs in live cells. Here, we use the malachite green aptamer (MGA), an RNA aptamer that can specifically bind to malachite green (MG) dye and induces it to emit far-red fluorescence signals. Previous studies on MGA showed a potential for the use of MGA for genetically tagging other RNA molecules in live cells. However, these studies also exhibited low fluorescence signals and high background noise. Here we constructed and tested RNA scaffolds containing multiple tandem repeats of MGA as a strategy to increase the brightness of the MGA aptamer-fluorogen system as well as to make the system fluoresce when tagging various RNA molecules, in live cells. We demonstrate that our MGA scaffolds can induce fluorescence signals by up to ~20 fold compared to the basal level as a genetic tag for other RNA molecules. We also show that our scaffolds function reliably as genetically-encoded fluorescent tags for mRNAs of fluorescent proteins and other RNA aptamers.

studies have shown that aptamer fluorescence in bacteria was only barely above cellular autofluorescence level, possibly due to high degradation and improper folding (15). One strategy to enhance the fluorescence signals was to stabilize MGA by encasing it within a 5S RNA (MGA-5S). One study showed preliminary findings where MGA fluorescence was seen in-vitro and in E. coli using MGA-5S (16). MGA-5S fluorescence in E. coli was further explored in another study where an inducible MGA-5S model showed up to a 3-fold increase in batch culture and up to a 10fold when cultured in a chemostat (15). This study also showed a concurrent expression of MGA and GFP fluorescence using a fusion RNA only when certain linkers were used to separate MGA-5S and the GFP mRNA (15).
Studies (8,9) have shown that adding many RNA repeats significantly enhanced fluorescence in bacterial and yeast cells using the MS2 systems. This approach of using RNA repeats has also been applied to enhance aptamer fluorescence using Spinach aptamer (17). It was seen that expression of a scaffold or array of multiple Spinach aptamer copies in bacterial cells enhanced Spinach-DFHBI (3,5-difluoro-4-hydroxybenzylidene imidazolinone) fluorescence as detected by imaging. Spinach repeats showing a 11-fold increase (17). Therefore, expressing a scaffold of multiple MGA repeats in tandem can, in principle, solve the issue of low fluorescence and high background invivo. A recent study using malachite green aptamer scaffold also showed increased fluorescence in-vitro but not in-vivo in mammalian and yeast cells (18).
Here, we explore the expression of a scaffold consisting of 6 copies of MGA in bacterial cells, showing a compatible spectrum of fluorescence with other currently available RNA aptamers. We observe a strong and stable MGA fluorescence that can be linked to multiple kinds of RNA molecules without significant fluorescent signal loss in E. coli. This MGA fluorescence can be studied using microscopy and flow cytometry showing its potential as a tool to study RNA expression and dynamics in-vivo.

MGA fluorescence is enhanced by aptamer scaffolds
We designed scaffolds of MGAs with various repeats that are separated by small linkers. The MGA scaffolds were integrated into a pET28a plasmid, which allowed MGA's in-vitro and cellular expression to be regulated by the T7 promoter and lac operator system respectively. First, six repeats of MGAs (MGA6) were designed by joining six MGAs with five distinct linkers, and then twelve repeats of MGAs (MGA12) were constructed by joining two different MGA6 scaffolds. MGA4 was produced by deletion of two different MGA structures. MGA-5S -a single MGA aptamer enclosed in a 5S ribosome subunit that has been used in prior work (15,16) -was also constructed in a pET28a plasmid. The plasmids were then transfected into BL21DE3 strain For all studies, MGA expression was induced by 0.1 mM Isopropyl β-D-1-thiogalactopyranoside (IPTG) and quantified immediately after the addition of 5µM MG. E. coli containing MGA6-pET28a and MGA12-pET28a, respectively, showed a very strong MGA fluorescent signal upon induction more than 20-fold higher than that of control ( Figure 1A). Cells containing MGA4-pET28a and MGA-5S-pET28a also showed a moderately strong MGA fluorescence upon induction, which was slightly smaller than that of MGA6 or MGA12. The distribution function of the MGA fluorescence signals at the single cell level was obtained ( Figure 1B) and the cell-to-cell variability in the fluorescence signals was quite strong (coefficient of variation for the induced sample =1.0). The induced MGA fluorescence was also seen by imaging using a microscope in the presence of MG ( Figure 1C).
The fluorescence signals of the MGA scaffolds can be enhanced further by optimizing cell density in the culture and by treating the culture with rifampicin. Our flow cytometry analysis has shown that the strength of the fluorescence signals per cell is positively correlated with cell density in the culture ( Figure S1). This may be related to the fact that the cells grow more slowly in dense culture and leads to slower dilution of the MGA caused by slower cell doubling, while MGA synthesis is kept roughly the same. Another way to enhance the signals is by rifampicin treatment. Rifampicin, an inhibitor of bacterial DNA-dependent RNA polymerase, does not inhibit the activity of T7 polymerase (22). Thus, it can make the transcription cellular machinery more dedicated to T7 polymerase activity. Figure 2 shows that MGA fluorescence was increased significantly after rifampicin treatment.
The fluorescence signals of the MGA scaffolds can be dynamically controlled by turning off the transcription of MGA. Removal of IPTG from the growth media by washing and suspending the cells in growth media without IPTG was seen to sharply decrease MGA fluorescence to the basal level within 1.5 hours (Figure 3), a time interval that corresponds to approximately two growth cycles ( Figure S2). This suggests that dilution due to cell doubling causes the reduction of the signal and more importantly that the MGA-MG signal half-lifetime is at least larger than the cell doubling time similar to that of Spinach (17).

Robust MGA fluorescence can be seen in the presence of fluorescent proteins
One advantage of MGA is its fluorescence spectrum at far-red wavelength that allows it to be used At the single cell level, we confirmed that a strong correlation between the levels of MGA and eYFP via flow cytometry ( Figure 4B) and microscopy ( Figure 4E). These results indicate the potential of using MGA as genetically encoded fluorescent tags for RNA molecules including in principle both mRNAs and regulatory RNAs to study their roles.

Robust MGA fluorescence can be seen in the presence of other aptamers
Recent in-vivo studies in E. coli involving the expression of aptamers such as Spinach and Mango have shown promising results (17,18,(23)(24)(25)(26)(27). The fact that fluorescence wavelengths from these aptamers are respectively green and yellow show that they could be used simultaneously with

Discussion
MGA is one of the first light-up aptamers to be described in the literature (12). However comparatively few studies have been conducted on MGA due to a variety of concerns that include high background to fluorescence ratio and cytotoxicity. Recent studies on other light-up aptamers such as Spinach (17), RNA Mango(25), Spinach 2(24), Broccoli(26), etc. have revealed various methodologies and techniques that can also be applied to the MGA system to overcome its shortcomings that will enable its use as a RNA labeling and imaging marker. Some of its potential advantages over widely applied RNA labeling methods such as using fluorescent protein-fused Previous studies have shown that MG concentrations below 10 µM do not have any effect on E.
coli viability, even though MG toxicity has been one of the concerns of using MGA in vivo (28). coli (17,23). This can be partly because the K d of MGA to MG dye (K d =0.11 µM) is lower than that of Spinach to DFHBI (K d =0.3 µM).
Previous studies on both Spinach(29) and MGA(15) have shown that placing an aptamer in either the 5' or 3' untranslated region (UTR) of a gene of interest decreases the fluorescent output of the aptamer possibly due to increased misfolding rates through interactions with proximal nucleotides.
However, using our MGA scaffolds we show that the fluorescence is not diminished in the presence of eYFP, eGFP, Spinach, or Mango.
A recent study on a similar scaffold of multiple MGA repeats in eukaryotic cells indicated that the MGA system can only be used for about 10 minutes after the addition of MG dye where the fluorescence/background levels remain sufficiently low (18). However, using MGA6 we observe a decreased yet detectable (approximately two-fold above the basal level) fluorescence even in 6 hours after the addition of MG to the growth media ( Figure S5). MGA fluorescence relies on the specific folding of MGA upon its binding to MG. It is to be noted that our MGA6 has different designs and different linkers (designed to maintain MGA folding) when compared with the MGA repeats used in the other study (18) which might impact RNA folding processes differently.
One main advantage of MGA6 is that it is smaller (429 bp) than fluorescent proteins or 8 repeats of Spinach aptamer. The smaller MGA4 (297 bp) or MGA-5S (189 bp) have also shown to be capable of a moderately strong fluorescence. A recent paper identified new analogs of MGA that are capable of binding to MG and induce fluorescence. One such analog, MGA 1.3 (26 bp) is smaller in size than the 38 bp core MGA that we used here. This analog can be potentially used to design MGA scaffolds to decrease size (30).
In conclusion, our scaffold of multiple MGA repeats was shown to be capable of generating strong MGA-MG fluorescence in E. coli under various growth conditions. This strong MGA fluorescence was not diminished when fused with fluorescent protein mRNAs or RNA aptamers. This study demonstrates an enhanced fluorescence quality of the MGA scaffold that can potentially be used for live cell studies using flow cytometry and microscopy. In addition, it will enable dual color monitoring of multiple RNA molecules, which can be used for direct understanding of RNA regulation and transcription-translation.

Plasmid construction
A scaffold of 6 MGA repeats (MGA6) was designed using the MGA sequence described in the literature (15). The MGA repeats were separated by randomized 17 nt linker sequences and two complementary base pair sequences at both ends of MGA that stabilize the MGA structure.
Predicted RNA folding determined by RNAfold (19) and Kinefold (20) was used in the design process to select for linkers that do not disrupt MGA native folding.
The MGA6, MGA12 and MGA-5S sequences were synthesized as double-stranded DNA and was delivered in the pUC57 plasmid (Genscript) between the XbaI and HindIII restriction sites (pUC57-MGA6) by Genscript. The scaffolds were excised with XbaI and HindIII and inserted into pET28a plasmid between the XbaI and HindIII restriction sites to generate pET28a-MGA6.
The MGA6 sequence was amplified pUC57-MGA6 using inserted PCR amplification and Phusion DNA polymerase (NEB). A pGA1A3 plasmid containing a synthetic biology circuit consisting of a plac promoter, a ribosome binding site (RBS), fluorescent protein and a terminator was similarly linearized using PCR amplification using Phusion DNA polymerase with primers that share overlap sequences with the MGA6 amplicon. The MGA6 and a backbone amplicons were ligated in the Gibson assembly method by using a one-step isothermal DNA assembly mixture that includes Phusion DNA Polymerase, T4 Ligase (NEB), and T5 Exonucleases (NEB). pGA1A3 is a high copy number plasmid with ampicillin resistance and a pMB1 origin of replication. The resulting plasmid DNA was transformed into chemically competent Turbo cells (NEB) and plated on selective LB agar.

Flow cytometry
Cells were inoculated into M9 media and grown overnight at 37°C and 250 r.p.m. shaking. From this seed culture, the cells were then diluted to a 1:10000 ratio and incubated at 37°C grown to OD600=0.05-0.1 in a shaker. MGA fluorescence was measured in a Sony Biotechnology ec800 flow cytometer using a 665 nm filter (equipped with a 660 nm splitter) and a 640 nm excitation laser. 50,000 events were collected for each sample per run and gated by using a 2-D normal distribution using python package FlowCytometryTools. A Medium Flush cycle was used after each run to avoid inter-sample contamination.

Fluorescence microscopy
Imaging was performed using a Nikon Ti-E inverted wide-field fluorescence microscope with a large-format scientific complementary metal-oxide semiconductor camera (NEO, Andor Technology) and controlled by NIS-Elements. Samples were kept at 30°C throughout the imaging process using an environmental chamber. The total imaging time for each experiment was 10 hours during which both brightfield and fluorescent images were captured every 3 minutes. Image processing and analysis was completed using ImageJ (NIH) (21).
Prior to measurement, cells were suspended in warm M9 media supplemented with 5 µM malachite green. 5 µl of suspended culture was placed on 0.5% M9 agar containing 5 µM malachite green.
A glass microscope slide was placed over the agar pad and sealed airtight with a 1:1:1 mixture of Vaseline, lanolin, and paraffin.

Supporting Information
Supporting Note: Five supplementary figures and DNA sequences of our MGA scaffolds.