A modular platform for bioluminescent RNA tracking

Abstract

A complete understanding of RNA biology requires methods for tracking transcripts in vivo. Common strategies rely on fluorogenic probes that are limited in sensitivity, dynamic range, and depth of interrogation, owing to their need for excitation light and tissue autofluorescence. To overcome these challenges, we report a bioluminescent platform for serial imaging of RNAs. The RNA tags are engineered to recruit light-emitting luciferase fragments (termed RNA lanterns) upon transcription. Robust photon production is observed for RNA targets both in cells and in live animals. Importantly, only a single copy of the tag is necessary for sensitive detection, in sharp contrast to fluorescent platforms requiring multiple repeats. Overall, this work provides a foundational platform for visualizing RNA dynamics from the micro to the macro scale.

Fig. 1. An optimized platform for tracking RNA dynamics.

A Strategy to visualize transcripts using RNA lanterns. The lanterns comprise fusions of MS2 coat protein (MCP) and PP7 coat protein (PCP) with NanoBiT fragments (SmBiT and LgBiT, respectively). Transcription of bait RNA (comprising MS2 and PP7 aptamers) drives NanoBiT heterodimerization. In the presence of furimazine (Fz), light is produced (λ_max = 460 nm, blue glow). The bicistronic construct encoding the RNA lantern is shown below the scheme. MCP and PCP were fused with HA and FLAG tags, respectively, for expression analyses. B Modeling of the RNA lantern complex. Crystal structures of MS2 (purple, 1ZDI), PP7 (orange, 2QUX), and NanoLuc (5IBO), highlighting LgBiT (dark blue) and SmBiT (cyan), were modeled in ChimeraX. The MCP-SmBiT/PCP-LgBiT complex was modeled by aligning the corresponding N- and C- termini of each protein and aligning the 5ʹ and 3ʹ ends of the aptamers. C Predicted secondary structure of flexible RNA bait, as calculated by RNAfold. D RNA lantern with a flexible RNA bait via an in vitro transcription and translation (IVTT) assay. The total flux observed ± RNA bait. P = 4.16 × 10⁻⁵ (95% confidence interval) calculated by a two-sided t-test. E Engineered rigid RNA baits provided robust photon output in IVTT assay. RNA baits containing varying spacers between the MS2 and PP7 aptamers were constructed, with spacer length (X nt, blue, denoted below each bar graph). Fold change in signal over RNA lantern alone is plotted. MS2, PP7, and M-3-P^mut denote baits comprising isolated aptamers or a mutated PP7 aptamer (Supplementary Fig. 3A, B), respectively, none of which were expected to result in RNA lantern assembly. Representative luminescence images are shown below the graph. F Rigid RNA baits comprising various MS2 and PP7 stem lengths. Luminescence readouts were acquired following IVTT. Fold change in signal over RNA lantern alone is plotted. Representative luminescence images are shown below the graph. Data are presented as mean values ± SD for n = 4 replicates (D, E, and F). Source data for (D, E, and F) are provided as Source Data files.

Fig. 2. Biochemical validation of RNA lanterns.

A The designer RNA bait comprises a 4-way-junction of MS2, PP7, and Ni-NTA aptamers. This unit assembles the RNA lantern components (MCP-SmBiT and PCP-LgBiT, fused to HA and FLAG epitopes, respectively). B Bioluminescent output from RNA bait (M-3-P) or inactive mutant (M-3-P^mut) combined with lysate from cells stably expressing the RNA lantern. C Fold change in bioluminescent signal over no-RNA controls. Each bar represents the mean of n = 3 replicates with dots showing data from individual replicates. D Affinity purification of the RNA lantern complex using the Ni-NTA aptamer. Various concentrations of RNA bait were used, and captured complexes were eluted using metal chelator (EDTA). E Bioluminescent output of lantern complexes captured using anti-HA or anti-FLAG conjugate-agarose beads. F Fold change in signal from lantern complexes retrieved in the presence of M-3-P bait versus no RNA from (E) P (95% confidence interval) values were calculated using a two-sided t-test. BL bioluminescence, arb.u. arbitrary units. Data are presented as mean values ± SD for n = 3 replicates. Source data for (B–F) are provided as Source Data files.

Fig. 3. Kinetics of RNA lantern assembly.

A Bioluminescent output from RNA bait (M-3-P, red), inactive mutant (M-3-P ^mut, black), or no bait (gold). Samples were mixed with lysate from cells expressing RNA lanterns and analyzed over time. Two replicate experiments were performed for each condition and all runs are shown. B Real-time RNA lantern assembly with purified RNA bait, along with RNAs produced via in vitro transcription. Lanterns were expressed by in vitro transcription and translation in reticulocyte lysate prior to addition of either purified M-3-P RNA bait (1 nM), M-3-P DNA (1 nM; 69-nt transcript), or GFP-M-3-P DNA (1 nM; 880-nt transcript). Luminescence was measured every 30 s, and mean luminescence values (from n = 3 replicates) are plotted. Data were normalized at the 35-min timepoint for each 30 s interval. Solid curves are data fit to the mono-exponential function $y=1-e^{-k\left(t-t_0\right)}$, where t₀ represents a delay due to RNA production. Luminescence is detectable within 30 s of RNA introduction and half-maximum signal is observed in 4 min. Source data for (A and B) are provided as Source Data files.

Fig. 4. Robustness and modularity of structured RNA baits.

A Comparison of M-3-P to a 12-copy unstructured RNA bait. Top: Schematic of the structured M-3-P RNA bait (single copy). Bottom: Schematic of a flexible RNA bait comprising 12 copies of the MS2 and PP7 aptamers (MP_12X). B MP_12x was evaluated against M-3-P at equimolar RNA concentrations (1 nM of template DNA) P = 6.10 × 10⁻¹³ or at 1/12 the concentration (83 pM of template DNA) P = 4.77 × 10⁻⁹. Fold change over no-RNA bait samples is plotted. Data are presented as mean values ± SD. C RNA bait assembles other lanterns. Schematic of MCP-PCP probes fused to split firefly luciferase (Fluc). MCP is fused to the N-terminal half of Fluc and PCP is fused to the C-terminal half. RNA transcription followed by D-luciferin treatment enables photon production and visualization (λ_max = 560 nm, green glow). D The Fluc lantern was assessed with a panel of RNA baits. Total light output for n = 6 replicates is shown. Only two replicates are shown for the no-RNA bait sample; the others were below the limit of detection represented by the electronic noise of the EMCCD camera (dashed line). Data are presented as mean values ± SD. E Graph of fold change in signal over no-RNA bait samples from (D) using a two-sided t-test. Data are presented as mean values ± SD. F Schematic of BRET-based RNA lanterns. G Luminescence fold change measurements for yellow and red RNA lanterns across a panel of RNA baits. Fold change calculated over no-RNA controls. P = 4.05 × 10⁻⁵ for yellow M-5-P vs M-7-P. Data are presented as mean values ± SD. H Emission spectra for BRET-based RNA lanterns. Error bars represent the standard deviation (SD) for n = 12 replicates in (B), n = 6 replicates in (D and E), n = 10 replicates in (G), and n = 3 replicates in (H). arb.u., arbitrary units. The P values (95% confidence interval) in (B, E, and G) were calculated by a two-sided t-test. Source data for (B, D, E, G, and H) are provided as Source Data files.

Fig. 5. Dynamic imaging in mammalian cells with RNA lanterns.

A Schematic of model mRNA encoding a fluorescent protein. The sequence was engineered with M-3-P in the 3′ UTR. Transcription of XFP-M-3-P mRNA results in lantern assembly and light emission (blue glow). XFP production can be further analyzed via fluorescence (green glow). (B) HEK293T cells expressing the RNA lantern were transiently transfected with DNA encoding GFP–M-3-P (Pearson’s r = 0.84) or GFP alone (Pearson’s r = 0.03). Luminescence was observed exclusively in cells containing mRNAs with the M-3-P bait. Scale bar = 20 µm. C Bulk measurement of photon flux from lantern-expressing cells transfected with DNA encoding GFP or GFP–M-3-P. P = 2.55 × 10⁻⁵ (95% confidence interval) calculated by a two-sided t-test. Data are presented as mean values ± SD for n = 3 replicates (D) Dynamic mRNA imaging. HEK293T cells expressing RNA lanterns were transfected with mCherry-β-actin (with M-3-P located in the 3′ UTR) and GFP-G3BP1. GFP-G3BP1 is known to localize to stress granules^,. Fluorescence readouts (mCherry) confirmed successful reporter transfection and expression. Cells were treated with sodium arsenite and imaged before (0 min) and after treatment (60 min), and n = 6 trials were conducted. Magnified views of two cell clusters are also shown. Scale bars = 20 μm. Source data for (B, C, and D) are provided as Source Data files.

Fig. 6. Imaging in live mice with RNA lanterns.

A Ai9 mice implanted with cells expressing RNA lanterns and GFP alone (left flank) or GFP–M-3-P (right flank) and imaged with luciferin. B Photon flux from Ai9 mice with implanted cells expressing GFP(±)M-3-P and RNA lanterns shown in (A). C Ai9 mice implanted with cells expressing RNA lanterns and BFP–M-3-P^mut (left flank) or BFP–M-3-P (right flank). D Photon flux from Ai9 mice with implanted cells expressing either BFP–M-3-P or BFP–M-3-P^mut and RNA lanterns shown in (C). Data are presented as mean values ± SD for n = 3 replicates (Fig. 6B, D). P values (95% confidence interval) were determined by unpaired, two-tailed t-test. BL bioluminescence. arb.u. arbitrary units. Source data for (A–D) are provided as Source Data files.