Sequence design
Nanostars were designed using NUPACK53 on the basis of published in vitro results11. For each design, ten NUPACK trials were run, and the one that generated the lowest defect score was selected. Broccoli, Pepper, Mango and MS2 aptamer sequences were taken from literature18,20,34,54. The wild-type KL (5′-GCGCGC) was adapted from the HIV-1 palindromic KL sequence32. Orthogonal KLs 5′-UCGCGA, 5′-GUCGAC and 5′-GGUACC were taken from the study by Fabrini et al.12. KLs 5′-GUAUAC and 5′-UAUAUA were designed by simply replacing GC pairs with AU pairs. Non-palindromic KLs were adapted from the 3sβ set designed by Stewart et al.11. Detailed design principles of nanostar stems and KLs can be found in Supplementary Section 1.1. All sequences are listed in Supplementary Table 1.
RNA synthesis for in vitro characterization
All RNA strands for in vitro experiments were transcribed from custom DNA templates synthesized by Integrated DNA Technologies as LabReady resuspensions with standard desalt purification. We annealed non-coding DNA templates with a 21-nt complement including the T7 promoter region and a 4-nt sealing domain (5′-GCGC). These templates were annealed in 1× TE/50 mM NaCl from 90 °C to room temperature at −1 °C min−1 at 5 μM for storage and used at 0.01 μM during in vitro transcription. RNA strands were transcribed in vitro at 37 °C using 7.5% (v/v) T7 polymerase from the AmpliScribe T7-Flash transcription kit (ASF3507, Biosearch Technologies), and transcription buffer prepared in-house: 40 mM of Tris-HCl, 10 mM of NaCl, 30 mM MgCl2, 2 mM spermidine, 7.5 mM of each NTP and 10 mM DTT.
Plasmid development
Inserts were directly purchased from Integrated DNA Technologies as two single-stranded, 5′-phosphorylated oligonucleotides containing the sequence of interest, flanked by NotI and SacII restriction sites. The two strands were annealed in 50 mM NaCl and 1× TE buffer using a heat treatment protocol including a 5-min melt at 90 °C, followed by a slow temperature ramp at −1 °C min−1, and held at 20 °C. The resulting products were double-stranded DNA fragments with sticky ends ready for ligation. After annealing, strands were purified with a DNA cleanup kit (NEB, T1030). The DNA encoding the nanostar sequences was inserted in the pAV-U6+27-Tornado-Broccoli (Addgene, 261587) plasmid. Plasmids were prepared by (1) digestion with NotI-HF (NEB, R3189S) (2 μl for 20-μl reactions) at 37 °C for 1 h, (2) purification with the DNA cleanup kit, (3) digestion with SacII (NEB, R0157S) (2 μl for 20-μl reactions) at 37 °C for 1 h and (4) purification with a 0.8% 1× TAE agarose gel to select the product with the correct size. Digested backbones were finally purified using a gel extraction kit (Qiagen, 28704). Digested backbone and inserts were ligated at a 1:10 molecular ratio by overnight incubation with T4 DNA ligase (NEB, M0202S) at 4 °C. Ligated plasmids were transformed into 50 μl DF5Hα competent cells (Thermo Fisher, EC0112 and 18258012) following the manufacturer’s protocol. We then extracted plasmid DNA using a Miniprep kit (Qiagen, 27106) following the manufacturer’s protocol. Extracted plasmids were finally sequenced by Eurofins Genomics (whole plasmid sequencing service). The plasmid expressing MCP-mCherry was purchased from Addgene (207668)55.
Cell culture and maintenance
HEK293T (ATCC, CRL-3216), HeLa (ATCC, CCL-2) and U-2 OS (ATCC, HTB-96) cells were grown in Dulbecco modified Eagle’s medium (DMEM), high glucose, pyruvate (Thermo Fisher, 11995065) containing 10% fetal bovine serum and 100 U ml−1 penicillin–streptomycin (Thermo Fisher) and maintained at 37 °C with 5% CO2 in a humidified incubator. Cells used for imaging were cultured in µ-Slide 8-well high slides (Ibidi GmbH).
Transfection
Seeding density was adapted across cell types to achieve ~70% confluence at transfection. Lipofectamine 2000 (Thermo Fisher, 11668019) was used for transfecting HEK293T cells. FuGene HD (Promega, E2311) was used for transfecting HeLa and U-2 OS cells as it demonstrated less cytotoxicity due to transfection. For experiments involving the expression of multiple nanostars, the total amount of plasmid DNA used in each experiment was kept constant, with an equal proportion of each nanostar variant.
Total RNA extraction
Cells were transfected in 24-well plates, as described in Supplementary Section 1.5. We changed the medium 24 h after transfection, and collected cells 48 h after transfection. For collection, we aspirated media, washed with PBS and trypsinized the cells. After trypsinization, cells were resuspended in PBS, lysed and RNA was purified using the Monarch Total RNA Miniprep Kit (NEB, T2010S). RNA concentration was estimated using a Nanodrop 2000c by measuring absorption at 260 nm.
Live-cell staining
The culture medium from overnight incubation was aspirated and replaced with fresh medium supplemented with two drops of NucBlue Live reagent (Hoechst 33342 nuclear dye, Thermo Fisher, R37605) per millilitre of media, along with the appropriate staining dyes according to experimental conditions. For conditions involving the Broccoli aptamer, we used 40 µM of DFHBI (Lucerna; 400-5 mg). For experiments involving the Pepper aptamer, we supplied 10 nM of HBC620 (MedChemExpress, HY-133520). Live cells were then incubated for at least 15 min at 37 °C before imaging. Cells were imaged in the presence of dyes.
Fixed cell staining and immunostaining
Mouse anti-coilin (Cajal body colocalization) was purchased from Abcam (ab11822; 1:1,900, 1 μg ml−1). Mouse anti-SC35 (nuclear speckle colocalization) was purchased from Abcam (ab11826; 1:200, 5 μg ml−1). Mouse anti-fibrillarin (nucleolus colocalization) was purchased from Antibodies.com (A85370; 1:200). Mouse anti-G3BP1 (stress granules colocalization) was purchased from Thermo Fisher (66486-1-IG; 1:200, 5 μg ml−1). Mouse anti-DCP1A (P-body colocalization) was purchased from Novus Biological (H00055802-M06; 1:200).
Before fixation, cell culture media were removed and cells were rinsed with PBS (Thermo Fisher, 10010023). Cells were then fixed in a PBS buffer (Thermo Fisher, 14190144) containing 4% paraformaldehyde (Thermo Fisher, 043368.9M) for 10 min at room temperature, and washed with the PBS buffer three times, each for 5 min. Next, we permeabilized cells with 0.5% Triton X-100 (Sigma-Aldrich, 9002-93-1) in PBS buffer for 10 min and washed three times. For imaging condensates involving the Mango aptamer, we added PBS supplemented with NucBlue reagent (Thermo Fisher, R37605) and 200 nM of TO1-B (ABM, G955). Cells were incubated in the buffer for 15 min before imaging. For experiments involving immunostaining, cells were further blocked using 3% bovine serum albumin (BSA) (w/v; Sigma-Aldrich, 9048-46-8) in PBS buffer for 1 h, and washed three times. Then, cells were stained with corresponding primary antibodies diluted to the above-mentioned concentrations with 3% BSA in PBS buffer and incubated at 4 °C overnight. The next day, primary antibodies were removed and cells were washed three times before the addition of secondary antibodies (Thermo Fisher, A-21236; 1:1,000 in 3% BSA in PBS buffer). We incubated cells in secondary antibodies for 1 h before removing the buffer and washing them three times with PBS. For the final wash, PBS was supplemented with 40 μM DFHBI and NucBlue reagent. Cells were incubated in the buffer for 15 min before imaging.
Microscopy
FRAP and fusion experiments were performed with epifluorescence imaging using a Nikon Eclipse TI-E inverted microscope and a 60× oil immersion objective. z-Stack confocal images were acquired using a Nikon Ti microscope equipped with an NL5+ camera. Images in Fig. 4d,k were captured using a Yokogawa CSU X1 spinning disk confocal on an inverted Zeiss stand. Hoechst (NucBlue staining) signals were detected in the UV channel (excitation 405 nm). Broccoli aptamer fluorescence was measured using the GFP channel (excitation 488 nm). Mango aptamer fluorescence was measured using the YFP channel (excitation 514 nm). Pepper aptamer, CY3 and mCherry fluorescence was detected using the RFP channel (excitation 561 nm). Finally, Alexa Fluor 647-labelled secondary antibody fluorescence was detected using the 647 nm channel (excitation 647 nm).
Image processing
To provide a more comprehensive view of condensate signals across all planes, confocal micrographs in the manuscript figures are max-pixel-intensity z-projections, unless otherwise specified in the figure caption. Detailed data processing methods and schematics can be found in Supplementary Information, including condensate volume and number quantification, nucleus volume quantification, partition coefficient calculation for linked condensates (JR and JG), partition coefficient calculation for peptides and small-molecule dye, and the PCC and Manders’ overlap coefficient (M1 and M2) calculation.
FRAP
FRAP experiments were performed using a Nikon Eclipse TI-E inverted microscope with a temperature control unit. In vitro samples were loaded into a house-made chamber and sealed with epoxy (Gorilla, 5-min set) for imaging. For in vivo experiments, cells were stained and imaged in the Ibidi chamber described above. Condensates were bleached with a 488-nm laser for 200 ms. For in vitro samples, imaging was captured once before bleaching and every 5 s for 10 min after bleaching. For in vivo samples, imaging was captured once before bleaching and every 200 ms for 2 min after bleaching or every 1.5 s for 5 min after bleaching. Images were analysed by extracting time-dependent average intensities within the bleached area and unbleached area. Data processing and fitting details are described in Supplementary Information.
Time-dependent coalescence analysis
For in vitro experiments, RNA strands were transcribed, labelled with 1% CY3-UTP, diluted ten times with transcription buffer and sealed in a chamber with epoxy, following the same protocol as FRAP experiments. Samples were imaged every 5 min for the first 4 h, then every 20 min until 10 h. We monitored condensate fusion events using a Nikon Eclipse TI-E inverted microscope with a temperature control unit. The temperature was maintained at 37 °C for all experiments.
For in vivo fusion experiments, we imaged cells (in media supplemented with 40 μM DFHBI and 2 drops ml−1 NucBlue) every 5 min for 60 min under the confocal microscope. The temperature was maintained at 37 °C. Fusion events were identified manually. Data processing was performed using a script in Python3, as described in our previous work11.
Flow cytometry
Flow cytometry experiments were performed using a BD FACSAria flow cytometer. Cells were transfected in 24-well plates, as described in Supplementary Section 1.5. We changed the media 24 h after transfection, and collected cells 48 h after transfection. For collection, we aspirated media, washed with PBS and trypsinized the cells. After trypsinization, cells were resuspended in PBS supplemented with 10% fetal bovine serum and dyes and filtered through a 40-μm cell strainer (Fisher Scientific, cat. no. 22363547) for flow cytometry. Hoechst was detected using a laser with Ex. 405 nm and a 450/50-nm filter; DFHBI was detected using a laser with Ex. 488 nm and a 530/30-nm filter.
RT–qPCR
Reverse transcription was carried out with equal amounts of RNA using the Protoscript II First Strand cDNA Synthesis Kit and random hexamers (New England Biolabs). Reverse-transcription quantitative PCR (RT–qPCR) was then performed using tenfold diluted cDNA and the Luna Universal qPCR Master Mix (New England Biolabs) in the CFX Real-Time PCR system (Bio-Rad), courtesy of the UCLA Virology Core. The qPCR conditions used were as previously described56,57. Target transcript levels were determined by normalizing the cycle threshold value of the target transcript to that of the housekeeping gene RPS11 transcript. Fold change was calculated using this normalized value relative to Lipofectamine control expression levels. For RT–qPCR primers, see Supplementary Table 4.
Polyacrylamide gel electrophoresis
Gel pre-mix was prepared by adding 42 g of urea to nanopure water, the mixture was then heated until the urea completely dissolved. This mixture was allowed to cool to room temperature, and then a 40% (v/v) 19:1 acrylamide:bis-acrylamide solution was added in the appropriate volume for the desired percentage (final volume of 100 ml). To start polymerization, 8 ml of pre-mix was added in appropriate ratios with TBE and nanopure water, ammonium persulfate and tetramethylethylenediamine. Gels were cast in 8 cm × 8 cm, 1-mm-thick disposable mini gel cassettes (Thermo Scientific, NC2010) and allowed to polymerize for 30 min before electrophoresis. After curing, the gel was pre-run in a 1× TBE buffer for 30 min. Wells were washed carefully to remove excessive urea. Samples and a low-range ssRNA ladder (NEB, N0364S) were prepared by mixing individual strands with denaturing RNA loading dye (NEB, B0363S), then heated at 70 °C for 10 min and immediately placed on ice. Owing to the low expression of exogenous RNA in mammalian cells, 5 μg of total RNA extraction was loaded into each well. Gels were run at room temperature at 100 V in 1× TBE unless otherwise noted. After electrophoresis, the gels were washed three times, each for 5 min, with nanopure water, then stained with DFHBI-1T staining buffer (10 μM DFHBI-1T, 40 mM HEPES, 100 mM KCl and 1 mM MgCl2) for 15 min. After staining, gels were imaged using the Bio-Rad Gel Imaging Systems. Then, gels were washed three times, each for 5 min, to remove the DFHBI-1T and stained in 1× SYBR Gold Nucleic Acid Gel Stain for 15 min and imaged again.
Statistics and reproducibility
At least three independent biological replicates were tested for each condition. For microscopy imaging, at least three different fields of view were captured for each replicate. No data were excluded from the analyses.
