Tandem Spinach Array for mRNA Imaging in Living Bacterial Cells

Live cell RNA imaging using genetically encoded fluorescent labels is an important tool for monitoring RNA activities. A recently reported RNA aptamer-fluorogen system, the Spinach, in which an RNA aptamer binds and induces the fluorescence of a GFP-like 3,5-difluoro-4-hydroxybenzylidene imidazolinone (DFHBI) ligand, can be readily tagged to the RNA of interest. Although the aptamer–fluorogen system is sufficient for imaging highly abundant non-coding RNAs (tRNAs, rRNAs, etc.), it performs poorly for mRNA imaging due to low brightness. In addition, whether the aptamer-fluorogen system may perturb the native RNA characteristics has not been systematically characterized at the levels of RNA transcription, translation and degradation. To increase the brightness of these aptamer-fluorogen systems, we constructed and tested tandem arrays containing multiple Spinach aptamers (8–64 aptamer repeats). Such arrays enhanced the brightness of the tagged mRNA molecules by up to ~17 fold in living cells. Strong laser excitation with pulsed illumination further increased the imaging sensitivity of Spinach array-tagged RNAs. Moreover, transcriptional fusion to the Spinach array did not affect mRNA transcription, translation or degradation, indicating that aptamer arrays might be a generalizable labeling method for high-performance and low-perturbation live cell RNA imaging.

To obtain the excitation and emission spectrum of Spi-tRNA, Spi and Spi-nR RNA, the following instrument parameters were applied: (1) excitation spectrum, emission wavelength 505 nm and emission slit width 5 nm, excitation wavelength 400 -480 nm and excitation slit width 5 nm; (2) emission spectrum, excitation wavelength 460 nm and excitation slit width 5 nm, emission wavelength 480 -580 nm and emission slit width 5 nm.

Fluorescence enhancement efficiency measurement
We defined the fluorescence enhancement efficiency as the fluorescence intensity increment when the aptamer repeat number is doubled in the Spinach array. The value of the enhancement efficiency is located between 0, when the fluorescence intensity remains unchanged, and 1, when the fluorescence intensity is doubled, upon aptamer number plication. The fluorescence intensity of Spi-nR, I Spi-nR , as a function of the enhancement efficiency can be formulated as: In the equation I Spi-nR is the in-vitro fluorescence of Spi-nR ( Figure 1C), I Spi is the in-vitro fluorescence of Spi ( Figure 1C), "eff" is the enhancement efficiency, and log 2 (n) is the value of logarithms base 2 of n, which we further defined as duplication round, or the number of repeat duplication required to construct Spi-nR from Spi. We can transform the formula into: log 2 (I Spi-nR ) = log 2 (1+eff) × log 2 (n) + log 2 (I Spi ) The enhancement efficiency value can be obtained by plotting log 2 (I Spi-nR ) as a function of log 2 (n) and conducting linear fitting to obtain the slope value, which is the log 2 (1+eff) in the equation. Then the enhancement efficiency can be calculated with the slope value. We put the data plot and linear fitting in Supplementary Figure S1, and had the slope value 0.67 and the fluorescence enhancement efficiency value 0.59.

Relative aptamer folding efficiency measurement of Spi-nR compared to Spi
The measure the relative folding efficiencies of Spi, Spi-8R and Spi-32R compared to Spi-tRNA, we kept the aptamer amount identical for Spi-tRNA, Spi, Spi-8R and Spi-32R, and incubated the RNA with excess amount of DFHBI. The solution was sent for fluorescence measurement and the fluorescence of Spi, Spi-8R and Spi-32R was compared with that of Spi-tRNA to calculate the relative folding efficiencies of Spi, Spi-8R and Spi-32R compared to Spi-tRNA. In the experiment we incubated 100 nM Spi-tRNA, 100 nM Spi, 12.5 nM Spi-8R or 3.125 nM Spi-32R with 10 µM DFHBI, respectively, making the aptamer concentration in each solution 100 nM and DFHBI 100 times the concentration of the aptamer.

Spinach binding kinetics measurement
The fluorescence increase was recorded with a fluorometer (Cary Eclipse Fluorescence Spectrophotometer, Agilent Technologies) after rapid mixing of certain concentration of Spinach RNA (40 nM Spi, 10 nM Spi-8R or 5 nM Spi-32R) with different concentrations of DFHBI (200 -700 nM) in a cuvette (100 μL). The excitation and emission wavelengths were 460 and 505 nm, respectively. The excitation and emission slit widths were 5 nm and 20 nm, respectively. The illumination intensity was estimated to be 1−5 W/cm2. We obtained the concentration of Spinach/DFHBI binding complex as a function of time (Supplementary Figure S11). We used MATLAB to simulate the time-course change of the complex concentration, and fit the simulated curve to our experimental data to determine k on and k off by maximizing R 2 : k on , k off and K D values are all reported in Supplementary  Table S2.The fluorescence intensity of Spi, Spi-8R and Spi-32R all displayed a biphasic behavior (Supplementary Figure  S11), a fast increase followed by a much slower increase. We hypothesized that additional conformational change of the aptamer resulted in a second slower phase with potentially altered k on and k off . However, with the single observable in our experiment, we cannot determine the parameters for the additional kinetic steps. Therefore, we focused on the fast phase, and simulated the first 100 s of the binding curve and determined the respective k on , k off and K D value of DFHBI binding onto Spi, Spi-8R and Spi-32R under various DFHBI concentration conditions (Supplementary Figure S12A). With the determined k on and k off values, we predicted a smaller value than the experimentally measured value at the plateau region. Nevertheless, the predicted value can capture 85% of the experimental value for Spi (Supplementary Figure S12B), and 80% for Spi-8R and Spi-32R.
We also found that for Spi, Spi-8R and Spi-32R, the k on , k off and K D value were consistent at various DFHBI concentrations (200 -700 nM). The average value of the k on and k off we calculated (Spi, k on = (8. (3.0 ± 0.5)× 10 -2 s -1 ) also indicated that the binding kinetics between DFHBI and the Spinach aptamer did not change significantly from single Spinach aptamer (Spi) to tandem Spinach arrays (Spi-8R and Spi-32R).

Time-resolved fluorescence measurement and fluorescence lifetime calculation
The fluorescence lifetime measurements were performed by a custom built confocal microscope as described elsewhere in details [Han et al., JACS (2013)]. An excitation light (473 ± 5 nm; ~ 10 ps pulse duration) was generated from an ultrafast laser (MaiTai HP, Spectra Physics) with the use of a photonic crystal fiber (FemtoWhite 800, NKT Photonics) and a pre-stretcher (N-SF57, CASIX). The fluorescence signal was detected by an avalanche photo diode (Micro Photon Devices) and registered by a time correlated single photon counting module (SPC630, Becker Hickl).

Bacteria growth and induction
E. coli cells were grown at 37 °C with antibiotics according to the plasmid selection markers (100 μg/mL ampicillin (Gold Biotechnology, Inc) for pUC57 and pUC57-Simple plasmid, 50 μg/mL Kanamycin (Roche Diagnostics) for pET28a and pET28c plasmid, 30 μg/mL Chloramphenicol (Sigma-Aldrich) for pTRUEBLUE-BAC2 plasmid, and 30 μg/mL Chloramphenicol for E. coli Rosetta strain) in Lysogeny Broth (LB) (LB Broth Miller, EMD Millipore) liquid and solid media. To gauge cell density, optical density (OD) of the medium was assayed at 600 nm using a plastic cuvette in a Spectramax Plus 384 Microplate Reader (Molecular Devices, Inc.). For RNA and protein induction, Rosetta cells transformed with a given plasmid were grown at 37 °C in LB medium overnight from a single colony, and diluted 1000fold into fresh LB medium and kept growing until at OD 600 = 0.2. IPTG (Sigma-Aldrich) was then supplemented with 1 mM final concentration to induce RNA and protein production under T7 promoter-lac operator control. To maintain exponential growth of the cells, pre-warmed medium was added to dilute the cell culture to OD 600 = 0.3 whenever OD 600 of the culture exceeded 0.5.

RNA decay assay
Cells were grown in LB medium and induced by IPTG as described above. After 60 min induction by 1 mM IPTG, the cells were centrifuged to a pellet and the supernatant was carefully aspirated and the cell pellet was resuspended in fresh pre-warmed LB medium without IPTG. The centrifugation and resuspension process was performed twice to remove the remaining IPTG, and the cell pellet was finally resuspended in LB medium without IPTG. The cells were grown at 37 °C, and at different time points after IPTG removal, a bit of cell culture was taken out for imaging.

Epifluorescence microscopy and image analysis
To prepare the imaging sample, 1 mL of cell culture was supplemented with DFHBI to 100 μM final concentration 10 min before imaging, and was kept growing at 37 °C for DFHBI permeation and binding to Spinach aptamer. The cells were then centrifuged and supernatant was removed. After cell resuspension in pre-warmed M9 minimal medium (M9, Minimal Salts, 5X; Sigma-Aldrich) supplemented with 2mM MgCl 2 and 100 µM DFHBI, a few µL were sandwiched between a glass coverslip (No. 1.5) and a thin slab of 1.5 % (w/v) agarose gel. M9 minimum medium containing 2mM MgCl 2 and 100 µM DFHBI or 1x PBS medium was used to dissolve the agarose and make the gel for live cell imaging or FISH imaging, respectively.

Pulsed illumination microscopy and image analysis
Pulsed illumination imaging was conducted by a home-built objective-TIRF microscope with an oil immersion objective (1.4 NA 100x, Olympus) equipped with an emCCD camera (iXon DU-887, Andor Technology). Illumination with a 473 nm laser (MLL-III-473, Opto Engine LLC) was controlled by a mechanical shutter (UniBlitz VMM-D3) through a National Instruments NI-6503 digital I/O controller card and synchronized to CCD via home-built software (cplc.illinois.edu/software) 4 , and the data acquisition and analysis procedures conducted. The recorded movie was processed by a MATLAB code to generate a collection of fluorescence images of consecutive single frames. The superposed image was conducted by stacking selected frames in ImageJ.

Total RNA extraction and purification
For each sample, the cell culture was measured OD 600 . We took out appropriate volume of the cell culture which contained the same total cell number of 1 mL 0.5 OD 600 cells. The cells were centrifuged at 5,000 g and 4 °C for 5 min, and the supernatant was carefully aspirated. The cell pellet was lysed by 1 mM lysozyme/TE buffer (10 mM Tris:HCl, 1 mM EDTA, pH = 8.0) and total RNA was extracted and purified from the cell lysate through RNeasy Mini Kit (Qiagen) according to the protocol. 10 µg extracted RNA were further treated with DNase using Turbo DNA-free kit (Life Technologies) to remove remaining DNA which interfered with qPCR experiments. Afterwards reverse transcription reaction was conducted with 100 ng RNA in a 20 µL reaction volume to synthesize cDNA required for qRT-PCR experiments using iScript cDNA Synthesis Kit (Bio-Rad) from the DNAse-treated RNA, according to the protocol.

Quantitative reverse transcription PCR (qPCR) and transcript number estimation
1 µL reverse-transcribed cDNA were taken out of the 20 µL total volume of each reverse transcription reaction, and were diluted to 10 µL. Regular PCR reactions were first conducted using the diluted cDNA and designed qPCR primers to confirm that proper cDNA products were generated. After that 1 µL of the diluted cDNA were supplemented with qPCR primer and reaction and detection SsoAdvanced SYBR Green Supermix (Bio-Rad) to a 20 µL reaction volume and qPCR reactions were assembled in a 96-well PCR plate (Bio-Rad). qPCR primers targeting the mRFP1-coding sequence were used to quantify RFP-Spi-nR RNA, and primers targeting 16S ribosomal RNA were used to quantify 16S ribosomal RNA as the internal standard. The qPCR reactions were conducted and monitored by Bio-Rad CFX96 Touch Real-Time PCR Detection System.
The expression level and the cellular transcript number of RFP or RFP-Spi-nR mRNA were roughly estimated by calculating the relative RNA expression level compared to that of 16S rRNA, using ∆C T Method 5 . We have made a series of known dilutions of the cDNA samples and created standard curves of mRFP1 and 16S rRNA qPCR primers by plotting the threshold-crossing cycle number (C T ) of the amplification curves, to estimate the amplification efficiencies of both primer pairs, and to confirm that their amplification efficiencies are similar and reliable (between 90% and 105%). The relative expression level of a specific sample between RFP (-Spi-nR) mRNA compared to that of 16S rRNA can be calculated simply with the primer amplification efficiencies extrapolated from the standard curves and the C T values measured by the qPCR experiments. The approximate RNA copy numbers per cell for RFP (-Spi-nR) mRNA were estimated by translating the relative value into absolute copy number using the value of 20,000 -70,000 16S rRNA molecules reported by previous study 3,6,7 .

RNA fluorescence in situ hybridization (FISH)
The FISH probes, which are DNA oligonucleotides with 3' amine modification, were designed and ordered from Biosearch Technologies, and combined and labeled with Cy5 NHS (GE Healthcare). The protocol of RNA FISH, including probe design and labeling, cell fixation and permeabilization, and probe hybridization, were reported by So and coworkers 8,9 . Cell preparation and sample hybridization are briefly described below.
After cell harvest (equivalent to 15 mL OD 600 = 0.4) and centrifugation (4 °C, 4500x g, 8 min), supernatant was removed and cells were resuspended in 1 mL freshly prepared fixation solution (1x PBS, 3.7% (w/w) formaldehyde) and gently shaken at room temperature for 30 min. Centrifuge (400x g, 8 min) the cell suspension, remove supernatant, and then wash twice in 1 mL 1x PBS. Resuspend the cells in 70% (w/v) Ethanol, and leave the cell suspension at room temperature for at least 1h to permeate the cell. Afterwards centrifuge (600x g, 7 min) and remove supernatant. Resuspend cells in 1 mL 10% wash buffer (10% formamide (v/v), 2x SSC) and leave at room temperature for a few minutes, and then centrifuge (600x g, 7 min) and remove supernatant. Resuspend the cells well in 50 µL hybridization solution, which is supplemented with proper amount of Cy5-labeled combined probes (15 ng for each probe binding sites according to the target RNA, i.e. 13 binding sites for RFP sequence, 24 binding sites for Spi-8R and 96 binding sites for Spi-32R) and 50 µL 10% hybridization buffer (10% dextran sulface (w/v), 10% formamide (v/v), 1mg/mL E. coli tRNA, 2x SSC, 0.2 mg/mL BSA, 2 mM Ribonucleoside Vanadyl Complex), and leave at 30 °C overnight. On the next day, take a few µL of hybridization sample, add 20 volumes of 10% wash buffer, followed by mixing and centrifugation (600x g, 7 min). Afterwards repeat the following steps 3 times: resuspend the cells with 20 volumes 10% wash buffer after removing supernatant, incubate for 30 min at 30 °C, and centrifuge and remove supernatant. Finally resuspend the cells in 1 volume of 2x SSC and the cell resuspension are ready to image. Prepare 1.5% (w/v) agarose/PBS gel for imaging sample preparation, as described before. Cell sample was imaged using epifluorescence microscopy with the filter set mentioned above for Cy5 FISH probe.

Supplementary Figures
Figure S1                        (black letters refer to linker sequence between "Spinach 2" aptamers; green letters refer to "Spinach 2" aptamer sequence) mRFP1 atggcctcctccgaggacgtcatcaaggagttcatgcgcttcaaggtgcgcatggagggctccgtgacacgagttcgagatcagagggcgagggccgcccctacgagggcac ccagaccgccaagctgaaggtgaccaagggcggccccctgcccttcgcctgggacatcctgtcccctcagttccagtacggctccaaggcctacgtgaagcaccccgccgac atccccgactacttgaagctgtccttccccgagggcttcaagtgggagcgcgtgatgaacttcgaggacggcggcgtggtgaccgtgacccaggactcctccctgcaggacgg cgagttcatctacaaggtgaagctgcgcggcaccaacttcccctccgacggccccgtaatgcagaagaagaccatgggctgggaggcctccaccgagcggatgtaccccgag gacggcgccctgaagggcgagatcaagatgaggctgaagctgaaggacggcggccactacgacgccgaggtcaagaccacctacatggccaagaagcccgtgcagctgcc cggcgcctacaagaccgacatcaagctggacatcacctcccacaacgaggactacaccatcgtggaacagtacgagcgcgccgagggccgccactccaccggcgcctaa T7 promoter-lac operator taatacgactcactataggggaattgtgagcggataacaattc (red letters refer to T7 promoter sequence, and green letters refer to lac operator sequence) lac promoter-lac operator tttacactttatgcttccggctcgtatgttgtgtggaattgtgagcggataacaatttc (blue letters refer to lacZYA promoter sequence, and green letters refer to lac operator sequence)