Clin Res Cardiol (2021) DOI DOI https://doi.org/10.1007/s00392-021-01843-w |
||
The circular RNA cZNF292 is conserved and alters endothelial cell shape and functions | ||
A. Heumüller1, A. Jones2, A. Mourão2, C. Shi3, I. Wittig4, A. Fischer1, C. Dieterich5, M. Potente3, T. Braun6, P. Grote1, N. Jaé2, M. Sattler7, S. Dimmeler8 | ||
1Institute of Cardiovascular Regeneration and Department of Cardiology, Goethe Universität Frankfurt am Main, Frankfurt am Main; 2Institute of Structural Biology, Helmholtz Zentrum München, Neuherberg; 3Angiogenesis & Metabolism Laboratory, Max-Planck-Institut für Herz- und Lungenforschung, Bad Nauheim; 4Functional Proteomics, Goethe Universität Frankfurt, Frankfurt; 5Klinik für Innere Med. III, Kardiologie, Angiologie u. Pneumologie, Universitätsklinikum Heidelberg, Heidelberg; 6Max Planck Institut für Herz und Lungenforschung, Bad Nauheim; 7German Research Center for Environmental Heath (GmbH), Helmholtz Zentrum München, Deutsches Forschungszentrum für Gesundheit und Umwelt (GmbH), Neuherberg; 8Zentrum für Molekulare Medizin, Institut für Kardiovaskuläre Regeneration, Goethe Universität Frankfurt am Main, Frankfurt am Main; | ||
Non-coding RNAs have been implicated in development, progression and treatment of cardiovascular diseases. Among these, circular RNAs (circRNAs) are a special case as they are mostly generated from protein-coding genes by back-splicing of the pre-mRNA. Although several circRNAs are regulated in cardiovascular disease in mice and humans, their functional role in vivo is rarely addressed. The lack of in vivo evidence for a functional role of circRNAs is mainly due to the difficulty to genomically delete circRNAs without altering linear host mRNA levels. To circumvent this issue, we identified locus conserved circRNAs that contain intronic elements, which can be genetically targeted for depletion without affecting the host mRNA. Of the top enriched circular RNA, we identified the pair of human cZNF292 and murine cZfp292 as prime target for functional characterization in the endothelium.
Employing CRISPR/Cas9 technology, we first validated efficient removal of the circRNA by genomic deletion of the intronic cassette of cZNF292/cZfp292 which we termed Exon1A in vitro observing specific deletion of cZfp292 in multiple clones. Similarly, systemic knockout mice were depleted of circular Zfp292 but did not express altered host gene levels. Importantly, phenotypical screening showed no influence of cZfp292 on postnatal vascular growth in the retina but revealed poor endothelial cell alignment within the thoracic aorta of cZfp292 mutant mice (Cell length/width: WT: 3.3±0.08 vs. KO: 2.43±0.05, p-value=0.0007). Mechanistically, we excluded a function as miRNA-sponge, translational template or transcriptional regulator. Instead, density gradient centrifugations discriminating between protein bound and free RNA showed that cZNF292 associated with proteins and proceeded to identify interacting proteins by RNA-pulldowns and mass spectrometry. Strikingly, the significant enriched protein Syndesmos (SDOS) known as interactor of the focal adhesion adapter protein Paxillin was previously shown to regulate primary cilia and to take part in cytoskeletal rearrangement by interaction with Syndecan-4. Of note, Sdc4-/- mice exhibit a similar poor flow alignment of the endothelium. We therefore verified the interaction between cZNF292 and SDOS in vitro by means of RNA-immunoprecipitations, revealing an interaction of SDOS with circular (Fold IgG: 6.57±1.8) but not linear ZNF292 (Fold IgG: 1.76±0.9). To address the functional overlap, we silenced cZNF292 and SDOS and observed a significant reduction of primary cilia and an altered endothelial morphology, accompanied by a more prone localisation of PXN in focal adhesions. To further understand the influence of cZNF292 on SDOS, we mapped the respective interaction sites. Of note, predicted SDOS binding sites clustered around the back-splice site and experimental exploration showed that the cZNF292/SDOS interaction is partly driven by the circRNA specific Exon1A (Ex4-oligo KD: 548 nM vs. Ex4+Ex1A KD: 288 nM). Vice versa, mutation of RNA-binding residues on SDOS abolished the interaction with cZNF292. Functionally, overexpression of this SDOS-binding mutant was sufficient to mimic the effects of cell alignment and PXN relocation upon cZNF292 silencing. Taken together, our revealed a role of the intronic spliced and conserved circRNA pair cZNF292/cZfp292 in the vasculature in vitro and in vivo, which influence endothelial cell shape. Mechanistically, cZNF292 acts by interaction with the protein SDOS downstream of SDC4 |
||
https://dgk.org/kongress_programme/jt2021/aP1581.html |