Targeted RNA condensation in living cells via genetically encodable triplet repeat tags

Abstract Living systems contain various membraneless organelles that segregate proteins and RNAs via liquid–liquid phase separation. Inspired by nature, many protein-based synthetic compartments have been engineered in vitro and in living cells. Here, we introduce a genetically encoded CAG-repeat RNA tag to reprogram cellular condensate formation and recruit various non-phase-transition RNAs for cellular modulation. With the help of fluorogenic RNA aptamers, we have systematically studied the formation dynamics, spatial distributions, sizes and densities of these cellular RNA condensates. The cis- and trans-regulation functions of these CAG-repeat tags in cellular RNA localization, life time, RNA–protein interactions and gene expression have also been investigated. Considering the importance of RNA condensation in health and disease, we expect that these genetically encodable modular and self-assembled tags can be widely used for chemical biology and synthetic biology studies.


Reagents and apparatus
DNA oligonucleotides used in this work were synthesized and purified by Integrated DNA Technologies (Coralville, IA) and W. M. Keck Oligonucleotide Synthesis Facility (Yale University School of Medicine).The detailed sequences have been listed in Table S1.PCR products were cleaned using a Monarch ® PCR & DNA Cleanup Kit [New England Biolabs (NEB), Ipswich, MA].All the RNAs for in vitro experiments were transcribed using a HiScribe™ T7 high-yield RNA synthesis kit (New England Biolabs, Ipswich, MA) and purified with G-25 columns.All chemicals were of analytical grade and obtained from Sigma or Fisher Scientific unless otherwise noted.All the concentrations of nucleic acids were measured with a NanoDrop One UV-Vis spectrophotometer.The gel electrophoresis was performed on a BioRad electrophoresis analyzer (Bio-Rad, Hercules, CA) and imaged on a Bio-Rad Gel Doc EZ imager.Fluorescence measurements in solution were conducted with a BioTek Synergy 2 fluorescence plate reader (λex= 485/20 nm, λem= 528/20 nm) (BioTek, Winooski, VT).Table S1.Sequences of RNAs used in this study.Sequences for the F30-2d×Broccoli, Pepper and near-infrared fluorescent protein (NirFP) are shown in green, red and magenta, respectively.The CAG or AC repeating sequences are bolded and the complementary sequences for transacting are underlined.

RNA name
Sequences (5'-3') Figure S1.(a) Fluorescence measurement as performed in solutions containing 4 μM corresponding RNA, 20 mM MgCl2, and 80 μM DFHBI-1T after annealing for the in vitro RNA condensate formation.F30-2d×Broccoli-tagged RNA strands containing 0×, 4×, 20×, 31×, 47×CAG repeats or 70×AC repeats are depicted as 0R, 4R, 20R, 31R, 47R, and 70AC, respectively.Shown are the mean and standard deviation (SD) values from three replicated experiments.(b) The correlation between RNA concentration and fluorescence signal intensities as measured in solutions containing 0-70 μM F30-2d×Broccoli RNA, 20 mM MgCl2, and 80 μM DFHBI-1T under the same imaging condition for the in vitro RNA condensation characterization.Shown are the mean and SD values from three replicated experiments.(c) Based on the calibration curve shown in the panel (b), the estimated RNA concentration within each condensate formed in Figure 1.Shown are the mean and the standard error of the mean (SEM) values from three representative images.Two-tailed student's t-test: ***, p<0.001; ns, not significant, p>0.05.

Figure S2 .
Figure S2.(a) In vitro fluorescence recovery after photobleaching (FRAP) measurement within condensates formed in solutions containing 4 μM 47R RNA, 20 mM MgCl2, and 80 μM DFHBI-1T.Green and red arrows indicate the bleached area and outer ring region, respectively.Scale bar, 1 μm.(b) Averaged fluorescence intensity as measured within the bleached area (green line) and the outer ring region (red line) and plotted over time for the FRAP measurement.Shown are the mean and SEM values from at least three representative condensate FRAP measurements.

Figure S3 .
Figure S3.In vitro kinetic measurement of the condensate formation.(a) The number of condensates per imaging view, (b) the partition ratio and (c) area of each condensate are plotted after a 3 min 95°C heating and 30 s ice cooling.The condensate solution contains 4 μM 47R RNA, 20 mM MgCl2, and 80 μM DFHBI-1T.The partition ratio is defined as the ratio of average fluorescence intensity inside individual condensate versus background signals in the solution.Shown are the mean and SEM values from at least three representative images, each imaging view equals to 4,430 μm 2 .

Figure S4 .
Figure S4.The phase diagrams of Mg 2+ and RNA concentration-dependent in vitro condensate formation.The measurements were performed in a solution after annealing for the condensate formation, which contains 10 mM Tris-HCl at pH=7.5, 100 mM KCl, and DFHBI-1T of 20-fold concentration as that of RNA.Blue (gray) dots represent the presence (absence) of >10 condensate with partition ratio >2 per 4,430 μm 2 imaging view from at least three replicated experiments.

Figure S5 .
Figure S5.In vitro characterization of CAG-repeat-mediated condensation of different scrambled RNAs.(a) The area and (b) partition ratio of each condensate is plotted as a function of the length of scrambled target RNAs.The measurement was performed in solutions containing 4 μM corresponding RNA, 20 mM MgCl2, and 80 μM DFH BI-1T after annealing for the condensate formation.Each data point represents one condensate.Shown are the mean and SD values from at least three representative images, each imaging view equals to 4,430 μm 2 .

Figure S6 .
Figure S6.In vitro characterization of CAG-repeat-mediated condensation of different target RNAs.(a) The area and (b) partition ratio of each condensate is plotted as a function of the length of CAG repeats for each target RNAs.The measurement was performed in solutions containing 4 μM corresponding RNA, 20 mM MgCl2, and 80 μM DFHBI-1T after annealing for the condensate formation.Each data point represents one condensate.Shown are the mean and SD values from at least three representative images, each imaging view equals to 4,430 μm 2 .

Figure S7 .
Figure S7.The cytotoxicity assessment of CAG-repeat-expressing bacterial cells.(a) Fluorescence imaging of BL21 Star TM (DE3) E. coli cells that express a pET-28c-F30-2d×Broccoli-0×, 20×, 31× or 47×CAG (0R, 20R, 31R or 47R) plasmid.These cells were incubated in DPBS buffer containing 1 μM SYTOX TM Blue for 2 h before imaging.As a positive control, one well of 0R cells were also treated with 1 mM tetracycline (TET) during the 2 h incubation with the SYTOX TM Blue dye.Scale bar, 3 μm.(b) Average cellular SYTOX TM Blue fluorescence as measured in individual cells.Each data point represents one cell.Shown are the mean and SD values.All the data is collected from at least three representative images.Two-tailed student's t-test: ***, p<0.001; ns, not significant, p>0.05.

Figure S8 .
Figure S8.The FRAP measurement in BL21 Star TM (DE3) E. coli cells that express a pET-28c-F30-2d×Broccoli-47×CAG plasmid (47R).(a, b) Representative FRAP images from five out of nine measured cells containing two condensates at opposite poles.All these five cells show the transfer of fluorescence signals between two poles after photobleaching.Blue arrows indicate the bleached pole.Scale bar, 1 μm.(c, d) Averaged fluorescence intensity from the bleached pole (green dot/line), unbleached opposite pole (red dot/line) and whole cell (grey dot/line) are plotted over time for the FRAP measurement.The insets illustrate the corresponding measured regions.

Figure S9 .
Figure S9.Mg 2+ concentration-regulated cellular RNA condensate formation.(a) Fluorescence imaging of BL21 Star TM (DE3) E. coli cells that express a pET-28c-F30-2d×Broccoli-0×CAG (0R) or pET-28c-F30-2d×Broccoli-47×CAG (47R) plasmid.These cells were incubated in DPBS buffer containing 0, 1, or 5 mM MgCl2 for 2 h before imaging.Scale bar, 2 μm.(b) The violin plot distribution of the number of foci per cell, as measured from 0R or 47R cells.Solid and dashed line indicate the median and interquartile value, respectively.The black cross indicates the mean value.(c) The averaged fluorescence intensity as measured in each individual cell.Each data point represents one cell.Shown are the mean and SD values.All the data is collected from at least three representative images.Two-tailed student's t-test: ***, p<0.001; ns, not significant, p>0.05.

Figure S10 .
Figure S10.Condensate formation-mediated changes in cellular RNA lifetime.(a) Fluorescence imaging over 24 hours of BL21 Star TM (DE3) E. coli cells that express a pET-28c-F30-2d×Broccoli-0×CAG (0R) or pET-28c-F30-2d×Broccoli-47×CAG (47R) plasmid.These cells were first IPTG-induced for 2 h and then after removing the IPTG, left in the DPBS buffer for different time before imaging.Scale bar, 3 μm.(b) The averaged fluorescence intensity as measured in each individual cell.Shown are box plots with min-to-max whiskers collected from at least three representative images.The top and bottom line, upper and lower box boundary, and inner line indicate the minimum and maximum data point excluding outliers, 75th, 25th percentile, and median of the data, respectively.(c) The violin plot distribution of the number of foci per cell, as measured from 0R or 47R cells after different time of incubation in the DPBS buffer.Solid and dashed line indicate the median and interquartile value, respectively.The black cross indicates the mean value.All the data is collected from at least three representative images.

Figure S11 .
Figure S11.Condensation of target RNAs in E. coli cells.(a) Fluorescence imaging of BL21 Star TM (DE3) E. coli cells that express a pET-28c vector encoding 0R-, 20R-, 31R-, or 47Rtagged OxyS, lacY or lacZ target RNAs.Scale bar, 5 μm.(b) The violin plot distribution of the number of foci per cell, as measured from corresponding cells.Solid and dashed line indicate the median and interquartile value, respectively.The black cross indicates the mean value.(c) The partition ratio of individual cellular foci as measured from the same batch of cells.Each data point represents one cellular condensate.Shown are the mean and SD values.All the data is collected from at least three representative images.Two-tailed student's t-test with Bonferroni correction: ***, p< 0.0003.

Figure S13 .
Figure S13.Trans-acting recruitment and regulation of target cellular RNAs.(a) Fluorescence imaging of BL21 Star TM (DE3) E. coli cells that express either Pepper-tagged 0× (0R-Pep) or 47×CAG-repeat RNA (47R-Pep).Scale bar, 5 μm.(b) Fluorescence imaging of BL21 Star TM (DE3) E. coli cells that express a pET-28c vector encoding Pepper-tagged near-infrared fluorescent protein (Nir-Pep) mRNA and a complementary strand-tagged 0R.Shown are the images before and after adding 1 μM HBC620 and 5 mM Mg 2+ in DPBS buffer.Scale bar, 2 μm.(c) Average cellular Pepper and (d) NirFP fluorescence signals as measured in 366 and 512 individual cells were plotted before and after adding 1 μM HBC620 and 5 mM Mg 2+ , respectively.Shown are violin plots with solid and dashed line indicating the median and interquartile value, respectively.The black cross indicates the mean value.Data was collected from at least three representative images.Two-tailed student's t-test: ***, p<0.001; ns, not significant, p>0.05.

Figure S14 .
Figure S14.RNA condensate formation in living P. aeruginosa cells that carry an inducible, chromosomally integrated T7 RNA polymerase.(a) Fluorescence imaging of ADD1976 P. aeruginosa cells that express a pET-28c vector encoding Pepper-tagged near-infrared fluorescent protein RNA (Nir-Pep) and F30-2d×Broccoli-tagged 0× or 47×CAG repeats with a complementary strand (0R-cN or 47R-cN).Images were taken after adding 1 μM HBC620 and 200 μM DFHBI-1T in DPBS buffer.Scale bar, 5 μm.(b) Percentages of cells that contain different numbers of foci.Shown are 100% stacked columns of ADD1976 P. aeruginosa cells that express 0R-cN/Nir-Pep or 47R-cN/Nir-Pep.White, light green, green and dark green indicate the percentage of cells that exhibit 0, 1-2, 3-4 and >4 foci, respectively.~13% of 0R-cN/Nir-Pep cells (95 in total) and 44% of 47R-cN/Nir-Pep cells (92 in total) contain at least one focus.(c) The partition ratio of each cellular foci as measured based on the Broccoli channel fluorescence levels.Shown are box plot with min-to-max whiskers.The top and bottom line, upper and lower box boundary, and inner line indicate the minimum and maximum data point excluding outliers, 75th, 25th percentile, and median of the data, respectively.Two-tailed student's t-test: ***, p<0.001.