The structural flexibility of RNA underlies fundamental biological processes, but there are no methods for exploring the multiple conformations adopted by RNAs in vivo. We developed cross-linking of matched RNAs and deep sequencing (COMRADES) for in-depth RNA conformation capture, and a pipeline for the retrieval of RNA structural ensembles. Using COMRADES, we determined the architecture of the Zika virus RNA genome inside cells, and identified multiple site-specific interactions with human noncoding RNAs.
Subscribe to Journal
Get full journal access for 1 year
only $20.17 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
All sequencing datasets have been deposited in ArrayExpress under accession number E-MTAB-6427. Base-pairing prediction, structure prediction, and clustering data are available in the Supplementary Data files. Source data for Figs. 1 and 3b are available online. Additional data that support the findings of this study are available from the corresponding authors upon request. A step-by-step protocol is available as a Supplementary Protocol and will be provided as an open resource in Protocol Exchange42.
Lu, Z. et al. Cell 165, 1267–1279 (2016).
Aw, J. G. A. et al. Mol. Cell 62, 603–617 (2016).
Sharma, E., Sterne-Weiler, T., O’Hanlon, D. & Blencowe, B. J. Mol. Cell 62, 618–626 (2016).
Helwak, A., Kudla, G., Dudnakova, T. & Tollervey, D. Cell 153, 654–665 (2013).
Ramani, V., Qiu, R. & Shendure, J. Nat. Biotechnol. 33, 980–984 (2015).
Sugimoto, Y. et al. Nature 519, 491–494 (2015).
Liu, Z.-Y. et al. eLife 5, e17636 (2016).
Alvarez, D. E., Lodeiro, M. F., Ludueña, S. J., Pietrasanta, L. I. & Gamarnik, A. V. J. Virol. 79, 6631–6643 (2005).
Friebe, P. & Harris, E. J. Virol. 84, 6103–6118 (2010).
Filomatori, C. V. et al. Genes Dev. 20, 2238–2249 (2006).
Pirakitikulr, N., Kohlway, A., Lindenbach, B. D. & Pyle, A. M. Mol. Cell 62, 111–120 (2016).
Watts, J. M. et al. Nature 460, 711–716 (2009).
Hahn, C. S. et al. J. Mol. Biol. 198, 33–41 (1987).
Manzano, M. et al. J. Biol. Chem. 286, 22521–22534 (2011).
Liu, Z.-Y. et al. J. Virol. 87, 6804–6818 (2013).
Akiyama, B. M. et al. Science 354, 1148–1152 (2016).
Rouskin, S., Zubradt, M., Washietl, S., Kellis, M. & Weissman, J. S. Nature 505, 701–705 (2014).
Ding, Y. et al. Nature 505, 696–700 (2014).
Spitale, R. C. et al. Nature 519, 486–490 (2015).
Guo, Y. E. & Steitz, J. A. Mol. Cell. Biol. 34, 3780–3787 (2014).
Chu, C., Qu, K., Zhong, F. L., Artandi, S. E. & Chang, H. Y. Mol. Cell 44, 667–678 (2011).
Kwok, C. K., Marsico, G., Sahakyan, A. B., Chambers, V. S. & Balasubramanian, S. Nat. Methods 13, 841–844 (2016).
Liu, Z.-Y. et al. J. Virol. 91, e00484-17 (2017).
Kwok, C. K., Ding, Y., Tang, Y., Assmann, S. M. & Bevilacqua, P. C. Nat. Commun. 4, 2971 (2013).
Schirle, N. T., Sheu-Gruttadauria, J. & MacRae, I. J. Science 346, 608–613 (2014).
Travis, A. J., Moody, J., Helwak, A., Tollervey, D. & Kudla, G. Methods 65, 263–273 (2014).
Langmead, B. & Salzberg, S. L. Nat. Methods 9, 357–359 (2012).
Markham, N. R. & Zuker, M. in Bioinformatics: Structure, Function and Applications (ed. Keith, J. M.) 3–31 (Humana Press, New York, 2008).
Saldanha, A. J. Bioinformatics 20, 3246–3248 (2004).
Lai, D., Proctor, J. R., Zhu, J. Y. A. & Meyer, I. M. Nucleic Acids Res. 40, e95 (2012).
Darty, K., Denise, A. & Ponty, Y. Bioinformatics 25, 1974–1975 (2009).
Gower, J. C. Biometrika 53, 325–338 (1966).
Love, M. I., Huber, W. & Anders, S. Genome Biol. 15, 550 (2014).
Smith, T., Heger, A. & Sudbery, I. Genome Res. 27, 491–499 (2017).
Dobin, A. et al. Bioinformatics 29, 15–21 (2013).
Li, H. et al. Bioinformatics 25, 2078–2079 (2009).
Lun, A. T. L. & Smyth, G. K. BMC Bioinformatics 16, 258 (2015).
Robinson, M. D. & Oshlack, A. Genome Biol. 11, R25 (2010).
Lun, A. T. L., Chen, Y. & Smyth, G. K. Methods Mol. Biol. 1418, 391–416 (2016).
Phipson, B., Lee, S., Majewski, I. J., Alexander, W. S. & Smyth, G. K. Ann. Appl. Stat. 10, 946–963 (2016).
Lun, A. T. L. & Smyth, G. K. Nucleic Acids Res. 42, e95 (2014).
Ziv, O. & Miska, E. A. COMRADES: crosslinking of matched RNAs and deep sequencing. Protoc. Exch. (in the press).
The authors thank A. Kohl (Centre for Virus Research, University of Glasgow, Glasgow, UK) and L.J. Pena and R.F. França (Fiocruz Recife, Pernambuco, Brazil) for providing the PE243 ZIKV RNA used to generate the virus stock. We thank Y. Galanty and F.M. Martínez for assisting with CRISPR–Cas9 knockout; T.D. Domenico and W. Matsushima for collapsing UMIs; G. Sanguinetti for suggesting the RNA-folding strategy; A.M. Sharkey and L. Gardner (University of Cambridge, Cambridge, UK) for providing the JEG-3 cells; S. Moss for assisting with risk assessment; M.S. Diamond, T. Sweeney, A. Firth, D. Jordan, A. Zeisel, and members of the Miska group for their comments; and C. Flandoli for illustrations. This work was supported by Cancer Research UK (C13474/A18583 and C6946/A14492 to E.A.M.; award no. A17197 to J.C.M.), the Wellcome Trust (104640/Z/14/Z and 092096/Z/10/Z to E.A.M.; 207507 to G.K. and M.M.G.; Senior Fellowship in Basic Biomedical Science 207498/Z/17/Z to I.G.), the Human Frontier Science Program (HFSP; LT000558/2015 to O.Z.), the European Molecular Biology Organization (EMBO; ALTF1622-2014 to O.Z.), the Blavatnik Family Foundation (postdoctoral fellowship to O.Z.), the UK Medical Research Council (G.K. and M.M.G.), EMBL (core funding to J.C.M.), NIGMS (grants R01GM104475 and R01GM115649 to I.J.M., L.F.R.G., and J.S.-G.), City University of Hong Kong (projects 9610363 and 7200520 to C.K.K.), the Croucher Foundation (project 9500030 to C.K.K.), the Hong Kong RGC (projects 9048103 and 9054020 to C.K.K.), the NSFC (Excellent Young Scientist Fund 81522025 to C.-F.Q.), and the Academy of Medical Sciences, UK (Newton Advanced Fellowship to C.-F.Q.).
The authors declare no competing interests.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Integrated supplementary information
a, Reverse transcription stalling assay indicating in vivo crosslinking positions in the human 5.8S rRNA from cells treated with the indicated psoralen derivative. Arrows indicate stalling events; Sequence is shown to the left. b, Bioanalyzer RNA profiles of ZIKV enriched or input RNA from cells inoculated with ZIKV or control non-inoculated cells. c, Enrichment of ZIKV RNA, or a control β-actin RNA measured by TaqMan PCR. Mean and s.d. of 4-5 biologically independent samples is shown. d-e, Non-cropped dot-blots showing enrichment of crosslinked RNA (d) or crosslink reversal (e). PD: pulldown; FT: flow through; Control: non-crosslinked sample; UVC: short wavelength UV. Experiments were repeated independently 3 times (a-e) with similar results.
a, In vivo detected interactions overlaid on the Ribovision human 18S phylogenetic ribosomal RNA secondary structure. Colour-code is indicative of the number of supporting chimeras for each base-pair. b, Precision and sensitivity of ribosomal RNA base-pairing detection by COMRADES. Mean and s.d. of 3 independent experiments are shown. Analysis is based on ZIKV enriched libraries, therefore the sensitivity of COMRADES is expected to be underestimated by this analysis.
Heat maps of ZIKV RNA-RNA interactions in crosslinked libraries and controls. Chimeras ligated in 5′-3′ and 3′-5′ orientations are plotted above and below the diagonal respectively. Experiments 1-3 represent independent experiments carried out at different days.
a, Arch plot representation of short and long-range RNA-RNA interactions along the ZIKV genome. Colours representing the number of non-redundant chimeric reads supporting each interaction. b, Distribution of RNA-RNA interactions by nucleotide distance between interacting chimeric partners. Left pie chart shows distribution of all interactions; right pie chart shows distribution of interactions that span less than 1,000 nucleotides.
a, Viewpoint histograms showing binding positions of the cyclization sequences along the ZIKV genome in control libraries; related to Fig. 2b. Viewpoint regions are marked by dashed red lines. b, The non-circular ZIKV genome conformation. Color code is indicative of the number of supporting chimeras for each base-pair. c, Heatmap of RNA-RNA interactions between the 5′ UTR and the envelope coding region. d, Viewpoint histogram showing binding of nucleotides at position 2-56 along the ZIKV genome. e, Newly identified 5′ UTR structure. Color code as described in (b).
a, Shannon entropy values calculated for each nucleotide along the ZIKV genome. Entropy may range from 0 to 13.4 bits; ZIKV coordinates are indicated by the position of genomic elements below. b, Inverse correlation between the degree of experimental support of base paired regions and their entropy. Pearson correlation coefficient values were calculated for each 1,000 structures. c, Shannon entropy values for a selected region of the ZIKV genome. d, Number of supporting chimeric reads for each base-pair shown in (c).
The ZIKV genome conformation with the highest chimeric-reads support. Color code is indicative of the number of supporting chimeras for each base-pair. *: Regions with exceptionally low Shannon entropy; Vertical lines indicate long-distance interactions.
Computationally predicted structures for ~1,000 nucleotides regions along the ZIKV genome, related to Fig. 2d. Each structure is plotted as a dot according to its folding energy (dG) and experimentally supporting evidence (chimera reads). Red squares indicating the structure with the lowest possible folding energy for each region. r: Pearson correlation coefficient.
Clustering of structures based on degree of similarity, related to Fig. 2e. Example of structures are shown with a color code representing the number of non-redundant chimeric reads supporting each interaction. Only regions demonstrating a clear clustering pattern are shown.
a-b, Co-clustering of non-shuffled and shuffled structures, related to Fig. 2e. Red scale indicates the number of chimeric reads supporting each non-shuffled structure (a). c, Representation of the percent of in vivo probed interactions in use in an ensemble of 5 structures per region, an ensemble of all 1,000 structures per region, or individual structures. Yellow lines represent mean and s.d. of 1,000 structures.
a-c, Site specific base-pairing between the ZIKV genome and small nuclear RNAs (a), tRNAs (b) and specific miRNAs (c) in COMRADES and controls.
(a), Electrophoretic mobility shift assay (EMSA) of 5′-labeled miR-21 (lanes 1, 2) upon addition of ZIKV 5′ CS (lane 2); and 5′-labeled ZIKV 5′ CS (lanes 3-5) upon addition of miR-21 (lane 4) or of Ago2-miR-21 complexes (lane 5). The experiment was independently repeated 3 times with similar results. b-h, In vitro measured affinities of Ago2-miR-21 to a wildtype ZIKV 5′ CS. Mean and s.e.m. of 3 independent samples are shown. (b), mutated 5′ CS (c-d), a fully seed matched mutated 5′ CS (e), mutated 5′ CS with no cHP stem-loop (f), extended 5′ CS sequence containing the DCS stem-loop (g), and a short target perfectly complementary to the miR-21 seed (h).
a, TaqMan PCR measurements of mature miR-21 expression levels in wildtype cells and CRISPR/Cas9 MIR21 deletion-clones. Values are normalized to spike-in control. b, Intracellular ZIKV RNA in MIR21 knockout and wildtype cells. Two-sided Student’s t-test p-values: ** =0.001; *** =0.0002, 4 degrees of freedom. c, Expression level of control and miR-21 psiCHECK-2 reporters upon treatment with miR-21 or control inhibitors. miR-21 expression values denote for Renilla/Firefly luminescence signals. d, Intracellular ZIKV RNA in miR-21 inhibited and control cells. Two-sided Student’s t-test p-value: *** =0.0003, 4 degrees of freedom. e, Replication of a ZIKV replicon carrying a wildtype 5′ CS or a 5′; CS - 3′ CS double mutant, pre-treated with miR-21 or non-targeting inhibitors. Two-sided Student’s t-test p-values: **** =3.5E-07; n.s.= 0.1 (non-significant), 10 degrees of freedom. wt: wildtype; KO: CRISPR-Cas9 MIR21 deletion-clones; cont: control. Mean and s.d. of 3 (a, b, d), 4 (c), and 6 (e) biologically independent samples are shown.
a-c, Intracellular levels of ZIKV envelope protein (ZIKV NS1), measured by FACS. Representative experiment out of 3 is shown. d-f, Gating strategy for the FACS measurements shown in (a-c). AF488: Alexa Fluor 488 labelled secondary antibody.
Supplementary Figures 1–14
Probing of RNA base-pairing using clickable psoralen
COMRADES detected base-pairing. COMRADES scores represent the number of chimeric reads supporting each base-pairing along the viral RNA. Base1 and base2 denote interacting nucleotide portions.
1,000 predicted structures for virus fragment 1
1,000 predicted structures for virus fragment 2
1,000 predicted structures for virus fragment 3
1,000 predicted structures for virus fragment 4
1,000 predicted structures for virus fragment 5
1,000 predicted structures for virus fragment 6
1,000 predicted structures for virus fragment 7
1,000 predicted structures for virus fragment 8
1,000 predicted structures for virus fragment 9
1,000 predicted structures for virus fragment 10
Structure clustering source data. Each predicted structure is assigned with clustering coordinates (X1 and X2), COMRADES score, and thermodynamic stability (dG).
About this article
Cite this article
Ziv, O., Gabryelska, M.M., Lun, A.T.L. et al. COMRADES determines in vivo RNA structures and interactions. Nat Methods 15, 785–788 (2018). https://doi.org/10.1038/s41592-018-0121-0
Science China Life Sciences (2020)
Frontiers in Cell and Developmental Biology (2020)
Journal of Biomedical Science (2020)
Nucleic Acids Research (2020)