A molecular beacon-based approach for live-cell imaging of RNA transcripts with minimal target engineering at the single-molecule level

Abstract

Analysis of RNA dynamics and localization at the single-molecule level in living cells has been predominantly achieved by engineering target RNAs with large insertions of tandem repeat sequences that are bound by protein-based or oligonucleotide-based fluorescent probes. Thus, individual RNAs are tagged by multiple fluorescent probes, making them detectable by fluorescence microscopy. Since large insertions may affect RNA processes including trafficking and localization, here we present a strategy to visualize single RNA transcripts in living cells using molecular beacons (MBs) - fluorogenic oligonucleotide probes - with minimal target engineering. The MBs are composed of 2'-O-methyl RNAs with a fully phosphorothioate-modified loop domain (2Me/PS_LOOP MBs), an architecture that elicits marginal levels of nonspecific signals in cells. We showed that MBs can detect single transcripts containing as few as 8 target repeat sequences with ~90% accuracy. In both the nucleus and the cytoplasm, mRNAs harboring 8 repeats moved faster than those with 32 repeats, suggesting that intracellular activities are less impeded by smaller engineered insertions. We then report the first MB-based imaging of intracellular dynamics and localization of single long noncoding RNAs (lncRNAs). We envision the proposed minimally-engineered, MB-based technology for live-cell single-molecule RNA imaging could facilitate new discoveries in RNA research.

Figure 1

Detection of single RNA transcripts harboring different tandem repeats of MB binding sites using anti-repeat MBs and smFISH. (a) Scheme of the 7 constructs expressing EGFP with tandem repeat sequences inserted upstream of the EGFP gene. The MB targets and the EGFP mRNA are transcribed as one molecule. After microporation of HeLa cells expressing EGFP mRNA harboring different numbers of MB targets with 5 µM anti-repeat MBs, the cells were fixed and permeabilized and smFISH was performed to assess the accuracy of MBs for detecting single RNA transcripts. Representative maximum intensity projection images of anti-repeat MBs (ATTO647N-labeled) and EGFP smFISH (TAMRA-labeled) in HeLa cells expressing EGFP engineered with (b) 32, 16, 8, or 4, and (c) 2, 1, or 0 tandem repeats of MB target sequence at 8 h are shown. (Scale bar, 10 µm) (d) Accuracy of anti-repeat MBs for the detection of single RNA transcripts engineered with different numbers of MB binding sites. A custom Matlab program was written to analyze the percentage of MB signals that colocalized with smFISH signals. Data represent mean ± SD of at least 10 cells. *Represents significant difference from the 32-repeats construct.

Figure 2

Diffusion kinetics of single EGFP mRNA transcripts harboring 32 or 8 MB target sequence repeats in the nucleus and the cytoplasm. HeLa-N1-32x and HeLa-N1-8x cells were microporated with 5 µM anti-repeat MBs, and time-lapse images were acquired 8 h after microporation (See Movies S1 and S2). Single particle tracking analysis was performed to determine the diffusion coefficient of individual RNAs as described in Materials and Methods. (a) The distribution of diffusion coefficients of single mobile pEGFP-N1-32x RNAs in the nucleus (n = 106 tracks) and in the cytoplasm (n = 295 tracks) analyzed from at least 23 cells. (b) The distribution of diffusion coefficients of single mobile pEGFP-N1-8x RNAs in the nucleus (n = 109 tracks) and the cytoplasm (n = 1067 tracks) analyzed from at least 33 cells. Inset shows the mean ± SE diffusion coefficients.

Figure 3

MB-based imaging of single NEAT1 transcripts in the nucleus. (a) Diffusion kinetics of single engineered NEAT1 lncRNA transcripts in the nucleus. HeLa-NEAT1-8x cells were microporated with 5 µM anti-repeat MBs, and time-lapse images were acquired 8 h after microporation (See Movie S6). Single particle tracking analysis was performed to determine the diffusion coefficient of individual RNAs as described in Materials and Methods. The distribution of diffusion coefficients of single mobile pNEAT1-8x RNAs in the nucleus (n = 200 tracks) were analyzed from 25 cells. Data representing immobile tracks (D_eff < 0.0006 µm²/s as defined by immobilized Tetraspek beads) are highlighted in white. Data representing mobile tracks are highlighted in black. 89% of the detected engineered NEAT1 transcripts are immobile. (b) Live-cell detection of NEAT1 colocalization with PSP1α proteins. 5 µM of anti-repeat MBs or control MBs were injected into HeLa-NEAT1-8x cells transfected with pEYFP-PSP1α constructs. Representative fluorescent images acquired within a few minutes post-injection are shown. (Scale bar, 10 µm).

Figure 4

MB-based imaging of single HOTAIR transcripts in the nucleus and the cytoplasm. (a) Images of anti-repeat MBs for the detection of single pHOTAIR-8x transcripts. Following microporation of NIH3T3 cells with or without pHOTAIR-8x with 5 µM of anti-repeat MBs, the cells were fixed and permeabilized and smFISH was performed to assess the accuracy of MBs for detecting single RNA transcripts. Representative maximum intensity projection images of MB and smFISH signals are shown. Note that in NIH3T3 cells not transfected with pHOTAIR-8x, lack of smFISH signal shows the specificity of the smFISH probes for the human HOTAIR transcript used. (Scale bar, 10 µm) (b) Diffusion kinetics of single pHOTAIR-8x lncRNA transcripts in HeLa cells. HeLa-HOTAIR-8x cells were microporated with 5 µM anti-repeat MBs, and time-lapse images were acquired 8 h after microporation. Single particle tracking analysis was performed to determine the diffusion coefficient of individual RNAs as described in Materials and Methods. The distribution of diffusion coefficients of single mobile pHOTAIR-8x RNAs in the nucleus (n = 113 tracks) and in the cytoplasm (n = 255 tracks) analyzed from 30 cells. Inset shows the mean ± SE diffusion coefficients.