Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Meta-Analysis
. 2017 Jan 5;65(1):25-38.
doi: 10.1016/j.molcel.2016.11.029. Epub 2016 Dec 22.

Distinctive Patterns of Transcription and RNA Processing for Human lincRNAs

Affiliations
Free PMC article
Meta-Analysis

Distinctive Patterns of Transcription and RNA Processing for Human lincRNAs

Margarita Schlackow et al. Mol Cell. .
Free PMC article

Abstract

Numerous long intervening noncoding RNAs (lincRNAs) are generated from the mammalian genome by RNA polymerase II (Pol II) transcription. Although multiple functions have been ascribed to lincRNAs, their synthesis and turnover remain poorly characterized. Here, we define systematic differences in transcription and RNA processing between protein-coding and lincRNA genes in human HeLa cells. This is based on a range of nascent transcriptomic approaches applied to different nuclear fractions, including mammalian native elongating transcript sequencing (mNET-seq). Notably, mNET-seq patterns specific for different Pol II CTD phosphorylation states reveal weak co-transcriptional splicing and poly(A) signal-independent Pol II termination of lincRNAs as compared to pre-mRNAs. In addition, lincRNAs are mostly restricted to chromatin, since they are rapidly degraded by the RNA exosome. We also show that a lincRNA-specific co-transcriptional RNA cleavage mechanism acts to induce premature termination. In effect, functional lincRNAs must escape from this targeted nuclear surveillance process.

Keywords: CPSF73; empigen; exosome; lincRNA; mNET-seq; phosphor CTD marks; polyadenylation; splicing; transcription termination.

Figures

None
Figure 1
Figure 1
Differential mNET-Seq Profiles for Protein Coding and lincRNA genes (A) mNET-seq strategy with each Pol II phospho-CTD modification color coded. (B and C) Color-coded heatmaps showing phospho-CTD profiles across individual (B) protein coding TUs and (C) lincRNA TUs ordered based on their transcription levels. Profiles are aligned to TSS and TES as indicated. Genes >1,000 nt (excluding some smaller protein coding and lincRNA genes) were divided into 100 bins. (D and E) (D) mNET-seq profiles across TARS (black for protein-coding gene) and (E) LINC01021 (green for lincRNA gene) using seven different Pol II antibodies as indicated. Gene maps show exons filled in and introns hatched. A chromatin-seq profile is run below the mNET-seq profiles. Blue reads are sense and red reads antisense transcripts. Reads per 108 mapped reads are indicated in brackets. See also Figure S1.
Figure 2
Figure 2
lincRNAs Are Inefficiently Spliced (A and B) (A) mNET-seq/S5P analysis of protein-coding gene PTCD3 and (B) lincRNA TUG1. HeLa cells were treated with Pla-B (red) or DMSO control (blue). Only sense transcripts are shown. (C) Meta-analysis across exon-intron junctions (5′ss) of annotated introns for protein-coding TUs versus lincRNAs. (D) Heatmaps for protein-coding versus lincRNA genes aligned to 5′ss −400 to +400 nt upstream and downstream. Percent of introns showing co-transcriptional 5′ss peaks is shown below, including all data repetitions, either with untreated, DMSO mock-treated, or Pla-B-treated HeLa cells. (E) pA+ and pA− NpRNA-seq profiles are shown for WDR13 versus lincRNA TUG1. (F) Splicing index from pA+ NpRNA-seq for protein-coding and lincRNA TUs (duplicates shown). See also Figure S2.
Figure 3
Figure 3
lincRNAs Are Largely Unpolyadenylated and CPA Independent (A and B) (A) mNET-seq/T4P analysis of GAPDH and (B) lincRNA TUG1. Vertical dotted line over GAPDH denotes PAS. (C) Meta-analysis of termination region (up to 7 kbp 3′ to TES) associated mNET-seq/T4P profiles, ±CPSF73 depletion by small interfering RNA (siRNA) treatment. siLuc indicates siRNA control treatment. Protein-coding TUs are shown on the left and lincRNA TUs on the right. (D) Quantitation of readthrough transcript levels following CPSF73 depletion characterized by GB-signal-normalized siLuc to siCPSF73 signal ratio in 10 kbp downstream of TES. (E) Gene-specific profiles (CDK9, histone H2A, histone H3, and LINC01021) for pA+ and pA−NpRNA-seq. (F) Quantitation of levels of pA−/pA+ transcripts for protein coding versus lincRNA TUs based on number of fragments overlapping TUs. Duplicate data are shown. (G) mNET-seq/T4P versus ChrRNA-seq profiles for MALAT1. mascRNA and PAS positions are indicated. (H) pA+/pA− RNA-seq for MALAT1. 3′ end of TU is expanded. See also Figure S3.
Figure 4
Figure 4
lincRNAs Are Chromatin Restricted and Degraded by the Nuclear Exosome (A) Transcription levels for coding, lincRNA, and antisense RNA in chromatin or nucleoplasm as well as exon numbers and gene lengths. (B) Density plots of chromatin and nucleoplasm fragments per kilobase of transcript per million of mapped reads (FPKM) levels (first 500 bp) for protein-coding and lincRNA TUs. (C) ChrRNA-seq versus NpRNA-seq for tandem lincRNA and LBR locus. (D) Density plots of FPKM levels in chromatin, nucleoplasm, and cytoplasm comparing protein-coding and lincRNA TUs. (E) Comparison of ChrRNA-seq, Np-RNA seq, and mNET-seq/total Pol II for lincPZP ± exosome (EXOSC3). lincPZP is antisense to the protein-coding gene PZP (not expressed in HeLa cells). (F) Quantitation of ratios of nucleoplasm to chromatin RNA levels for different classes of transcript as indicated. Non-coding RNA (ncRNA) denotes stable RNA, such as snRNA and snoRNA. See also Figure S4.
Figure 5
Figure 5
Identification of Co-associated RNA-Processing Complexes with Pol II Comparison mNET-seq/S5P, S2P, Y1P, and T4P profiles with or without Empigen treatment for (A) MYC, (B) MIR17HG, (C) MALAT1, and (C) LINC01021, respectively. Orange arrows denote Empigen-sensitive peaks. See also Figure S5.
Figure 6
Figure 6
Effect of DGCR8 Depletion on Co-transcriptional Processing mNET-seq/S5P profiles for (A) CCND1, (B) MIR17HG-GPC5, (C) MALAT1, and (D) LINC01021 with DGCR8 siRNA-mediated depletion or control siLuc treatment. Orange arrow indicates loss of pre-miRNA cleavage for MIR17HG or elevated levels of cleavage products following DGCR8 depletion for MALATI and LINC01021. Duplicate mNET-seq/S5Ps are presented to underline data reproducibility. See also Figure S6.
Figure 7
Figure 7
Protein Coding versus lincRNA Defining Features: PCA and Model (A) Principal-component analysis applied to protein-coding and lincRNA TUs shown separately and merged. Vectors indicating key parameters compared are shown by arrows: these are exosome sensitivity, pA−/pA+ levels, cytoplasmic/chromatin, and nucleoplasmic/chromatin levels. Some key lincRNAs are identified as well as some protein-coding transcript outliers. The graph has been cropped for better visualization, but PC1 and PC2 values of all data points are available in Table S2. (B) Model for protein-coding versus lincRNA co-transcriptional processing. Protein-coding genes are transcribed by Pol II with spliceosome (pink oblong) associated with CTD S5P (red dot). mRNA 3′ ends are generated co-transcriptionally by CPSF73 as part of CPA complex, which in turn promotes Pol II termination. lincRNA genes are weakly spliced and polyadenylated, resulting in CPSF73-independent termination and DGCR8-stimulated exosome degradation with co-transcriptional cleavage (scissors) associated with CTD S2P and S5P (orange dot) and exosome-mediated degradation on chromatin. See also Figure S7.

Similar articles

See all similar articles

Cited by 51 articles

See all "Cited by" articles

References

    1. Affymetrix ENCODE Transcriptome Project. Cold Spring Harbor Laboratory ENCODE Transcriptome Project Post-transcriptional processing generates a diversity of 5′-modified long and short RNAs. Nature. 2009;457:1028–1032. - PMC - PubMed
    1. Andersen P.R., Domanski M., Kristiansen M.S., Storvall H., Ntini E., Verheggen C., Schein A., Bunkenborg J., Poser I., Hallais M. The human cap-binding complex is functionally connected to the nuclear RNA exosome. Nat. Struct. Mol. Biol. 2013;20:1367–1376. - PMC - PubMed
    1. Andersson R., Gebhard C., Miguel-Escalada I., Hoof I., Bornholdt J., Boyd M., Chen Y., Zhao X., Schmidl C., Suzuki T., FANTOM Consortium An atlas of active enhancers across human cell types and tissues. Nature. 2014;507:455–461. - PMC - PubMed
    1. Boutz P.L., Bhutkar A., Sharp P.A. Detained introns are a novel, widespread class of post-transcriptionally spliced introns. Genes Dev. 2015;29:63–80. - PMC - PubMed
    1. Brown J.A., Bulkley D., Wang J., Valenstein M.L., Yario T.A., Steitz T.A., Steitz J.A. Structural insights into the stabilization of MALAT1 noncoding RNA by a bipartite triple helix. Nat. Struct. Mol. Biol. 2014;21:633–640. - PMC - PubMed

MeSH terms

LinkOut - more resources

Feedback