Conversion of adenosine to inosine is a frequent type of RNA editing, but important details about the biology of this conversion remain unknown because of a lack of imaging tools. We developed inoFISH to directly visualize and quantify adenosine-to-inosine-edited transcripts in situ. We found that editing of the GRIA2, EIF2AK2, and NUP43 transcripts is uncorrelated with nuclear localization and paraspeckle association. Further, NUP43 exhibits constant editing levels between single cells, while GRIA2 editing levels vary.
This is a preview of subscription content
Subscribe to Journal
Get full journal access for 1 year
only $9.92 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Tax calculation will be finalised during checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Bass, B.L. & Weintraub, H. Cell 48, 607–613 (1987).
Liddicoat, B.J. et al. Science 349, 1115–1120 (2015).
Higuchi, M. et al. Nature 406, 78–81 (2000).
Kumar, M. & Carmichael, G.G. Proc. Natl. Acad. Sci. USA 94, 3542–3547 (1997).
Chen, L.-L. & Carmichael, G.G. Mol. Cell 35, 467–478 (2009).
Zhang, Z. & Carmichael, G.G. Cell 106, 465–475 (2001).
Prasanth, K.V. et al. Cell 123, 249–263 (2005).
Raj, A., van den Bogaard, P., Rifkin, S.A., van Oudenaarden, A. & Tyagi, S. Nat. Methods 5, 877–879 (2008).
Levesque, M.J., Ginart, P., Wei, Y. & Raj, A. Nat. Methods 10, 865–867 (2013).
Seeburg, P.H., Single, F., Kuner, T., Higuchi, M. & Sprengel, R. Brain Res. 907, 233–243 (2001).
Yamashita, T. et al. Neurosci. Res. 73, 42–48 (2012).
Paschen, W., Hedreen, J.C. & Ross, C.A. J. Neurochem. 63, 1596–1602 (1994).
Melcher, T. et al. Nature 379, 460–464 (1996).
O'Connell, M.A., Gerber, A. & Keller, W. J. Biol. Chem. 272, 473–478 (1997).
Yoshida, M. & Ukita, T. Biochim. Biophys. Acta 157, 455–465 (1968).
Sakurai, M., Yano, T., Kawabata, H., Ueda, H. & Suzuki, T. Nat. Chem. Biol. 6, 733–740 (2010).
Wang, I.X. et al. Cell Rep. 5, 849–860 (2013).
Shaffer, S.M. et al. Lab Chip 15, 3170–3182 (2015).
Sunwoo, H. et al. Genome Res. 19, 347–359 (2009).
Mellis, I.A. et al. Protocol Exchange DOI: http://dx.doi.org/10.1038/protex.2017.056 (2017).
Duffy, D.J. et al. Oncotarget 6, 43182–43201 (2015).
Kuhlwilm, M., Davierwala, A. & Pääbo, S. PLoS One 8, e83218 (2013).
Whitney, N.P. et al. FASEB J. 22, 2888–2900 (2008).
Cabili, M.N. et al. Genome Biol. 16, 20 (2015).
We thank K. Nishikura and M. Sakurai for helpful discussions, and we thank C. Bartman and B. Emert for comments and suggestions. I.A.M. acknowledges support from NIH/NIGMS T32GM007170 (University of Pennsylvania MSTP), NIH/NHGRI T32HG000046 and NIH/NINDS F30NS100595; S.H.R. from NIH 1F32GM120929-01A1; and A.R. from NSF CAREER Award 1350601, NIH New Innovator 1DP2OD008514, NIH/NIBIB R33 EB019767, NIH 4DN U01 HL129998, and NIH Center for Photogenomics RM1 HG007743.
A.R. receives royalties and consulting income from Biosearch Technologies related to Stellaris RNA FISH.
Integrated supplementary information
(Top) Hybridization mix including all fluorescently labeled probes is added to cells. Colocalization of guide probes with detection probes determines sites of adenosine-containing or inosine-containing transcripts. (Bottom) Multiple, singly labeled probes target the transcript of interest flanking the editing site. Individual probes targeting the editing site are generated and labeled with 2 different fluorophores--these probes contain matching mask-binding regions. When the mask is bound, the region that binds to the target mRNA is shortened and a mismatch destabilizes the hybridization. The two detection probes compete for binding and the specific probe remains bound. The mask is removed by strand displacement and the full sequence hybridizes to its target.
(a) GRIA2, (b) EIF2 AK2 and (c) NUP43: schematic gene representation, 5’ to 3’ regardless of strand (top), genotyping by Sanger sequencing of genomic DNA PCR around locus encoding editing site in cell line of interest (middle), and RT-PCR/Sanger verification of RNA editing in cell line of interest (bottom).
(a) Most GRIA2 transcripts are retained in the nucleus in SH-SY5Y cells. (b) GRIA2 guide probe specificity assessed by odds-evens probe set division experiment. (c) Nuclear retention of GRIA2 upon treatment with aphidicolin, a cell cycle inhibitor. (d) Western blot of GluR2 protein in SH-SY5Y cell lysate.
InoFISH results for each replicate, including pixel-shift and dye-swap controls for GRIA2 in SH-SY5Y cells (top), EIF2AK2 in U-87 MG cells (middle) and NUP43 in U-87 MG cells (bottom).
(top) Schematic representation of the experiment. Control experiment is normal GRIA2 inoFISH with GRIA2 detection probes. On the right, the guide probe has been replaced to target SFPQ with the GRIA2 detection probes. (bottom) Fraction of colocalized guides with GRIA2 (left) and SFPQ (right).
Supplementary Figure 6 Resampled mean GRIA2 editing level estimates from ADAR2 knockdown and cyanoethylation experiments.
(a) ADAR2 transcript abundance upon siRNA-mediated ADAR2 knockdown (left); GRIA2 inoFISH results in SH-SY5Y cells (n = 2) after ADAR2 knockdown and scrambled siRNA control (right). (b) Cyanoethylation reaction schematic (left) and GRIA2 inoFISH (right) results in SH-SY5Y cells (n = 3) after in situ cyanoethylation with -acrylonitrile control. (c) Resampled mean editing levels of scrambled siRNA vs ADAR2 siRNA knockdown. (d) Resampled mean editing levels of samples treated by cyanoethylation and control samples (-acrylonitrile). 10000 resampled mean estimated editing level values per boxplot.
inoFISH mean +/- s.e.m after cyanoethylation treatment from a minimum of 2 biological replicates (left) including pixel-shift and dye-swap controls per replicate (right) for GRIA2 in SH-SY5Y cells (top), EIF2AK2 in U87 MG cells (middle), and NUP43 in U87 MG cells (bottom).
(a) Pipeline outline for RNA-seq-based screen and GRIA2 editing level estimation. (b) Experimental pipeline and representative Sanger sequencing trace for RT-PCR/Sanger-based estimate of GRIA2 editing level. (c) Experimental pipeline and representative Sanger sequencing traces from genotyping of RT-PCR/Topo TA cloning-based estimate of GRIA2 editing level. (d) Experimental pipeline and Bioanalyzer report for RT-PCR/BbvI-digestion-based estimate of GRIA2 editing level.
(top) Schematic representation of the experiment. Control experiment is normal GRIA2 inoFISH with GRIA2 detection probes. On the right, the guide probe has been rcolocalized with an adenosine detection probe and a cytosine detection probe that doesn’t target either nucleotide. (bottom) Fraction of colocalized GRIA2 guides by normal inoFISH (left) and inoFISH with false competition (right).
Boxplots of 10000 parametric bootstrapped samples of editing level per boxplot for GRIA2 (top), EIF2AK2 (middle), and NUP43 (bottom). Mean editing level models resampled irrespective of subcellular localization (left) and considering nuclear vs cytoplasmic localization (right).
Outline of experimental design pipeline for inoFISH, including representative results (LSG1 RT-PCR/Sanger and MYO1 C genotyping/Sanger) at critical steps in the target selection process.
About this article
Cite this article
Mellis, I., Gupte, R., Raj, A. et al. Visualizing adenosine-to-inosine RNA editing in single mammalian cells. Nat Methods 14, 801–804 (2017). https://doi.org/10.1038/nmeth.4332
Nature Biotechnology (2019)
Genome Biology (2018)