Profiling the long noncoding RNA interaction network in the regulatory elements of target genes by chromatin in situ reverse transcription sequencing

Long noncoding RNAs (lncRNAs) can regulate the activity of target genes by participating in the organization of chromatin architecture. We have devised a “chromatin-RNA in situ reverse transcription sequencing” (CRIST-seq) approach to profile the lncRNA interaction network in gene regulatory elements by combining the simplicity of RNA biotin labeling with the specificity of the CRISPR/Cas9 system. Using gene-specific gRNAs, we describe a pluripotency-specific lncRNA interacting network in the promoters of Sox2 and Pou5f1, two critical stem cell factors that are required for the maintenance of pluripotency. The promoter-interacting lncRNAs were specifically activated during reprogramming into pluripotency. Knockdown of these lncRNAs caused the stem cells to exit from pluripotency. In contrast, overexpression of the pluripotency-associated lncRNA activated the promoters of core stem cell factor genes and enhanced fibroblast reprogramming into pluripotency. These CRIST-seq data suggest that the Sox2 and Pou5f1 promoters are organized within a unique lncRNA interaction network that determines the fate of pluripotency during reprogramming. This CRIST approach may be broadly used to map lncRNA interaction networks at target loci across the genome.

. CRIST-seq assay to map the promoter lncRNA interactome. Chromatin-RNA in situ reverse transcription trap sequencing (CRIST-Seq) assay. Cells were transfected with CRISPR dCas9 gRNA to target the gene promoter. The Cas9-gRNA expressing cells were crosslinked by formaldehyde to fix the promoter RNA chromatin structure. After cell membrane lysis, the nuclei were isolated and the promoter-interacting RNAs were in situ reverse transcribed into cDNAs with biotin-dCTP. The promoter biotin-cDNA chromatin complex was first immunoprecipitated by a Cas9-FLAG antibody, and the promoter-interacting biotin-cDNAs were separated from genomic DNAs by streptavidin beads. The CRIST-captured chromatin cDNAs were aliquoted for Illumina library sequencing or quantitative PCR to measure the enrichment of the identified lncRNAs to the target gene promoter.
A. The CRISPR Cas9 Sox2-gRNA vector used in CRIST assay. Cas9: CRISPR Cas9; gRNA1, 2: two Cas9 guiding RNAs that target the Sox2 promoter (sequences under the diagram); pEF1: the human EEF1A1 promoter; pU6: U6 promoter; pH1: human H1 promoter; T5: the TTTTT termination signal of RNA polymerase III. B. Location of the two Cas9 gRNAs in the Sox2 promoter. The Sox2 exon 1 mRNA is shown in blue and the coding region with ATG in green. Two Cas9 gRNAs are highlighted in yellow and PAM sequences in red in the Sox2 promoter region. TS+1: transcription initiation site. C.
A. The CRISPR Cas9 Pou5f1-gRNA vector. gRNA1, 2: two Cas9 guiding RNAs that target the Pou5f1 promoter (sequences under the diagram). B. Specific CRIST targeting of the mouse Pou5f1 promoter. pPou5f1: the targeting site in the Pou5f1 promoter where the Cas9 gRNAs are designed; 5'-Ct: the Pou5f1 control site that is 13.9 kb away from the pPou5f1 target site. Cas9 Vector: cells that were treated with the Cas9 control vector that lacks the gRNAs; Cas9-gRNA: cells that were targeted by both Cas9 and Pou5f1 gRNAs; Cas9-gCT: cells that were treated with the random control gRNA vector. Offtarget: a CRIST control site that is 33.8 kb upstream of the housekeeping gene GAPDH. The chromatin complex was immunoprecipitated with a Cas9-FLAG antibody and an IgG control antibody. CRIST signals were quantitated by real-time PCR using specific primers derived from the pOct4 targeting site, 5'-Ct control site and off-target site. All data shown are mean ±SEM from three independent experiments by normalization over the IgG control. ** p<0.01 as compared with Vector and gCT controls. dCas9-FLAG pH1 gRNA1 T5 pU6 gRNA2 T5 Figure S4. Specific CRIST targeting of the human IGF2 imprinting promoters.
A. The CRISPR Cas9 IGF2-gRNA vector. gRNA1, 2: two Cas9 guiding RNAs that target the IGF2 imprinting promoters (P2-P4, sequences under the diagram). B. Specific CRIST targeting of the growth factor IGF2 promoter. The human IGF2 has four promoters and nine exons. The promoter P1 is not imprinted and it drives the biallelic expression of the growth factor from exons 1, 2, 3, 7, 8, and 9. In contrast, its promoters P2-P4 are imprinted and are expressed exclusively from the paternal allele. In human tumors, however, this imprinting mechanism is dysregulated and causes biallelic expression of the mitogen that promotes tumor growth. We used the CRIST assay to examine the imprinted P2 promoter. pIGF2: the site in the IGF2 promoter where the Cas9 gRNAs are designed; 5'-Ct: IGF2 control site that is 106 kb away from the pIGF2 target site. Vector: cells that were treated with the Cas9 control vector that lacks the gRNAs; Cas9-gRNA: cells that were targeted by both Cas9 and IGF2 gRNAs; Cas9-gCT: cells that were treated with the random control gRNA vector. Off-target: a CRIST control site that is 33.8 kb upstream of the housekeeping gene GAPDH. The chromatin complex was immunoprecipitated with a Cas9-FLAG antibody and an IgG control antibody. CRIST signals were quantitated by real-time PCR using specific primers derived from the pIGF2 targeting site, 5'-Ct control site and offtarget site. All data shown are mean±SEM from three independent experiments by normalization over the IgG control. ** p<0.01 as compared with Vector and gCT controls. dCas9-FLAG pH1 gRNA1 T5 pU6 gRNA2 T5 Figure S5. Specific CRIST targeting of the human FLI1 promoter.
A. The CRISPR Cas9 FLI1-gRNA vector. gRNA1, 2: two Cas9 guiding RNAs that target the FLI1 promoter (sequences under the diagram). B. Specific CRIST targeting of the oncogenic FLI1 promoter. pFLI1: the site in the FLI1 promoter where the Cas9 gRNAs are designed; 5'-Ct: FLI1 control site that is 18.5 kb away from the pFLI1 targeting site. Vector: cells that were treated with the Cas9 control vector that lacks the gRNAs; Cas9-gRNA: cells that were targeted by both Cas9 and FLI1 gRNAs; Cas9-gCT: cells that were treated with the random control gRNA vector. Off-target: a CRIST control site that is 33.8 kb upstream of the housekeeping gene GAPDH. The chromatin complex was immunoprecipitated with a Cas9-FLAG antibody and an IgG control antibody. CRIST signals were quantitated by real-time PCR using specific primers derived from the pFLI1 targeting site, 5'-Ct control site and off-target site. All data shown are mean±SEM from three independent experiments by normalization over the IgG control. ** p<0.01 as compared with Vector and gCT controls. A. The FECR1-FLI1 CRIST vector. FECR1 is a known circRNA that binds to the FLI1 promoter and regulates its activity in cis. We thus used it as a CRIST positive control. Two Cas9 FLI1 gRNAs are transcribed by U6 and H1 promoters, respectively, and guides the dCas9 to the FLI1 promoter. The FLI1 CRIST-seq library was used to quantitate the enrichment of FECR1 circRNA in the FLI1 promoter. B. Enrichment of FECR1 in the FLI1 promoter. Vector: cells that were treated with the Cas9 control vector that lacks the gRNAs; Cas9-gRNA: cells that were targeted by both Cas9 and IGF2 gRNAs; Cas9-gCT: cells that were treated with the random control gRNA vector; Vector: cells that were treated with the empty dCas9 vector. The enrichment of FECR1 in the FLI1 promoter was quantitated by qPCR. For comparison, the value of the IgG group was set as 1. ** p<0.01 as compared with the IgG control group. C. The nuclear lncRNA MALAT1 is used as the FLI1 CRIST negative control. The binding of MALAT1 to the FLI1 promoter was quantitated by qPCR and was standardized over the IgG group.     Figure S9. Differentially expressed Sox2 promoter-interacting RNAs.
The Sox2 CRIST-seq RNAs were integrated with the RNA-seq >2 fold RNAs data using a VENN program (http://bioinformatics.psb.ugent.be/webtools/Venn). A cut-off threshold of peak enrichment FPKM >50 was arbitrarily set to select CRIST-Seq RNAs. The RNAs are listed in the order of expression fold change of iPSCs over fibroblasts. The abundance of the Sox2 pre-mRNA in the Cas9 Sox2-gRNA CRIST-seq library was detected by two primer sets from the Sox2 coding sequence. The third pair of PCR primers (JH4813/JH4815) from the downstream sequence was used as the negative control. The iPSC cDNA was used as the positive PCR control. Using this more sensitive PCR, we detected the presence of Sox2 pre-mRNA in the CRIST-seq products. As expected, no signals were detected in the Cas9 gCT and the Vector control groups.  A. Snhg14 RNA-DNA FISH. The isolated iPSC clones were subjected to RNA FISH (green) using biotin-14-dCTP-labeled single strand DNA probes for Snhg14. RNA probes are designed to cover the intron region for specific staining of the mature RNAs. Slides were subsequently re-fixed and processed for DNA FISH (red) using digoxigenin-dUTP labeled BAC probe. The nuclear DNA was stained with DAPI (blue). Arrows mark the colocalization of the Snhg14 lncRNA signal with the Pou5f1 DNA signal. B. Palr35 RNA-DNA FISH. Although lncRNA Palr35 was also differentially expressed in reprograming, it was not in the list of the Sox2 CRIST-seq. We thus used it as the RNA FISH control. As the control, RNA-FISH did not detect the co-localization of Palr35 with Sox2.
A. The RAT assay. iPSCs were fixed and Snhg14 lncRNA was labelled by biotin-dCTP in the nucleus using four Snhg14-specific oligonucleotides. The random oligonucleotides were used as the RAT control. After sonication, the biotin-Snhg14-chromatin complex was pulled down by streptavidin beads. The chromatin DNA was purified for qPCR. B. Location of qPCR primers. Primers B and C are designed from the Sox2 promoter. C. The Snhg14-Sox2 interaction by qPCR. The Snhg14-pulled down chromatin DNAs were mapped by a series of primers in the Sox2 locus. Note the enriched signal of the Sox2 promoter (B, C sites) in the Snhg14-pulled down chromatin complex.