Mechanisms of Long Nocoding RNA Nuclear Retent.pdf

ReviewMechanisms of Long Noncoding RNANuclear RetentionChun-Jie Guo,1Guang Xu,1and Ling-Ling Chen1,2,3,*LongnoncodingRNAs(lncRNAs)are crucialregulatorsin diversecellular contextsandbiologicalprocesses.Thesubcellularlocalizationof lncRNAsdetermines theirmodes of action.Compared to mRNAs,however,many mRNA-like lncRNAs arepreferentially localized tothe nucleus where they regulate chromatin organization,transcription,and different nuclear condensates.Recent studies have revealedthe complex mechanisms that govern lncRNA nuclear retention.We reviewcurrentunderstanding of how the transcription and processing of lncRNAs,motifswithin lncRNAs,and trans-factors coordinately contribute to their nuclear reten-tion in mammalian cells.Many lncRNAs Are Localized to the NucleusLarge-scaletranscriptomic analyseshaveuncoveredthatlncRNAs arepervasivelytranscribedbyRNA polymerase II(Pol II)13.Most well-characterized lncRNAs are processed to producemRNA-like molecules that contain a 5 methyl guanylate cap(m7G)and a 3 poly(A)tail 3,4,while some are generated via unusual processing pathways and are stabilized by forming uniquestructures at their ends(reviewed in 5,6).Although some annotated lncRNA transcripts can in-deed be translated into functional polypeptides 79,an increasing list of non-protein-codinglncRNAs play key roles in diverse cellular contexts and biological processes(reviewed in1012).These activities are associated with their distinct subcellular fates(reviewed in 5,13).Although some lncRNAs are exported to the cytoplasm to carry out their functions(Box 1),manyothers are localizedinthenucleus3,4,14,wheretheyassociatewithchromatin to modulatechromatin status(Figure 1A)and interfere with Pol II transcription(Figure 1B,C),or accumulate innuclear condensates to regulate their assemblies and functions(Figure 1DG).Compared to mRNAs,how a group of mRNA-like lncRNAs are acquired their distinct sub-cellular fates or functions has remained intriguing.On the one hand,the export of mRNA-like lncRNAs may copy mRNA mechanisms 15 or display selective localization similarly tomRNAs involved in biological processes such as DNA repair,proliferation,and cell survival16.On the other hand,recent studies have begun to reveal different factors that controlthe nuclear localization of mRNA-like lncRNAs.In this review we highlight aspects oftranscription and processing,cis-motifs,and trans-factors that contribute to lncRNA nuclearretention in mammalian cells.Distinct Transcription and Processing of lncRNAs Leads to Nuclear RetentionAlthough exhibiting tissue-or cell type-specific expression,most annotated lncRNAs arecapped,polyadenylated,and spliced similarly to mRNAs 13,and lncRNAs and mRNAsmay thus share common export pathways including diverse export factors such as TREX(transcription export)and NXF1(nuclear export factor 1)/NXT1(nuclear transport factor 2-likeexport factor 1)15.A recent study has shown that NXF1 is preferentially required for theexport of single-or few-exon RNAs with long exons or high A/U content,including manylncRNAs 17.HighlightsMany lncRNAs are transcribed bydysregulated RNA polymerase II andare inefficiently spliced,leading to pre-ferred nuclear localization patterns.Exonic repeats,C-rich motifs,and U1motifs in lncRNAs promote their specificnuclear localization and function.Trans-factors such as hnRNPs,RNAhelicases,and splicing factors facilitatelncRNAnuclearlocalizationbyinteractingwith different cis-acting motifs.The subcellular localization of lncRNAorthologs is regulated during evolution.1State Key Laboratory of MolecularBiology,Shanghai Key Laboratory ofMolecular Andrology,Chinese Academyof Sciences,Center for Excellence inMolecular Cell Science,ShanghaiInstitute of Biochemistry and CellBiology,University of the ChineseAcademy of Sciences,CAS,Shanghai200031,China2School of Life Science and Technology,ShanghaiTech University,Shanghai201210,China3School of Life Science,HangzhouInstitute for Advanced Study,Universityof Chinese Academy of Sciences,Hangzhou 310024,China*Correspondence:(L.-L.Chen).Trends in Biochemical Sciences,November 2020,Vol.45,No.11https:/doi.org/10.1016/j.tibs.2020.07.001947 2020 Elsevier Ltd.All rights reserved.Trends in Biochemical SciencesAn official publication of the INTERNATIONAL UNION OF BIOCHEMISTRY ANDMOLECULAR BIOLOGYDespite the fact that mature lncRNAs are mRNA-like,lncRNA genes are less evolutionarilyconserved,less abundantly expressed,and contain fewer exons than mRNA genes 13,18(Figure 2A,B).A significant portion of lncRNAs are derived from chromatin regions that havedifferent modifications from mRNA genes 4,19,and are transcribed by dysregulated Pol II20,weakly spliced 4,14,19,20,inefficiently polyadenylated 19,20,and unstable 20(Figure 2B).Because transcription and processing are known to be coordinated with efficient(m)RNA export 16,21,22,such altered transcription and processing of lncRNAs may facilitatethe inefficient nuclear export of a subset of lncRNAs.Chromatin status is an important indicator for active or suppressive gene regulation,and histonemodifications are the principle components of chromatin regulation 23.Analysis of chromatinmarks of lncRNA and mRNA gene promoters showed that lncRNA promoters have elevated levelsof repressive modifications such as enrichment in histone H3 trimethylated on lysine 9(H3K9me3)4,24(Figure 2A,B).Further analysis of promoters of the expression-level matched lncRNAs andmRNAs showed that lncRNA promoters contain fewer transcription factor(TF)binding sites butare enriched in conserved TF binding sites for GATA2,KAP1,and MBD4 4.Features ofH3K9me3 and TF binding sites on lncRNA genes likely contribute to their tissue-specific expressionpatterns 4,but their correlation with the subcellular localization of lncRNAs has remained unclear.RNA transcription and processing also differ between lncRNA and mRNA genes(Figure 2A,B).Capturing different phosphorylated Pol II C-terminal domains(CTDs)with mammalian native elon-gating transcript sequencing(mNET-seq)25 showed that a significant fraction of lncRNAs aretranscribed by dysregulated Pol II(CTD S2P and S5P)20.Many lncRNA transcripts are weaklyspliced,non-polyadenylated,andaresimultaneouslydegradedbynuclearexosomesonchromatin20,indicating that chromatin-associated degradation of lncRNAs may increase the local concen-tration of lncRNAs at chromatin,resulting in their reduced distribution to nucleoplasm and cyto-plasm.By contrast,mRNAs are transcribed by Pol II(CTD S5P),spliced,and polyadenylated toproduce mature transcripts 20.Thus,in principle,functional lncRNAs must escape this targetednuclear surveillance process to accumulate to high levels in specific cell types.Further analyses combining gene architectures,chromatin marks,expression levels,splicingefficiency,and sequence elements of cytoplasmic and nuclear fractions of RNA-seq datasetsBox 1.Examples of lncRNAs That Have Functions in the CytoplasmlncRNAs in the cytoplasm can function as miRNA sponges,regulate mRNAs turnover,and modulate translation.linc-RoRis highly expressed in the cytoplasm of differentiated ESCs and sequesters miR-145 that in turn suppresses the miRNA-mediateddegradationofmRNAsforthecoretranscriptionfactorsOct4,Sox2,andNanog77(seeFigure1Hinmaintext).lncRNA-PNUTS is generated from pre-mRNA of PNUTS by alternative splicing that removes 61 nt in the 5 region of exon12,thus disrupting the open reading frame of PNUTS mRNA.This contains seven miR-205 sites and promotes epithelialmesenchymal transition(EMT)in the context of breast cancer cell migration and invasion 78(see Figure 1H in main text).ThelncRNAMACC1-AS1promotesgastriccancercellmetabolismbystabilizingMACC1mRNA79(seeFigure1Iinmaintext).SomelncRNAs are not evenlydistributed in the cytoplasm and are insteadlocalized to specific regions.Forexample,ZFAS1(ZNFX1 antisense RNA 1)80,H19(H19 imprinted maternally expressed transcript)81,and TUG1(taurineupregulated 1)81 bind to ribosomes and regulate ribosome production,assembly,and translation(see Figure 1J in maintext).RMRP(RNA component of mitochondrial RNA-processing endoribonuclease)is localized to the mitochondrial matrixbutisencodedbythenuclearDNA.RMRPbindstoHuR(humanantigenR)inthenucleusandisexportedintothecytoplasmvia a CRM1(chromosomal maintenance 1,also known as exportin 1)-dependent export pathway.Upon arrival in thecytoplasm,itistransported tomitochondriabyinteractingwithGRSF1(G-rich sequencefactor1)inthe matrix,whereRMRPis essential for maintaining oxygen consumption rates and mitochondrial DNA replication priming 82(see Figure 1K inmain text).SAMMSON(survival-associated mitochondrial melanoma-specific oncogenic non-coding RNA)is also primarilylocalized to mitochondria(see Figure 1K in main text)and controls mitochondrial homeostasis,metabolism 83,and rRNAmaturation 84 by interacting with p32,a protein that regulates mitochondrial 16S rRNA maturation,the expression ofmitochondrially encoded polypeptides,and oxidative phosphorylation 85,86.Trends in Biochemical SciencesAn official publication of the INTERNATIONAL UNION OF BIOCHEMISTRY ANDMOLECULAR BIOLOGY948Trends in Biochemical Sciences,November 2020,Vol.45,No.11have revealedthat many factors contributeto lncRNAnuclear retention,and thatinefficient splicingis likely to be a major contributor 19.Comparison of the splicing efficiency of RNA from cytoplas-mic and nuclear fractions in human and mouse cells found that cytoplasmic RNAs have higherTrendsTrends inin BiochemicalBiochemical Sciences SciencesFigure1.FunctionsofLongNon-CodingRNAs(lncRNAs)AreDictatedbyTheirDistinctSubcellularLocalizationPatterns.(A)Firre(functionalintergenicRNArepeating element)scaffolds nuclear architecturesacross chromosomes.hnRNPU(heterogeneous nuclear ribonucleoprotein U)is required forFirre trans-chromosomal localiza-tion via RRD(repeating RNA domain)on Firre 48,49.(B)Jpx(just proximal to Xist)blocks chromosome inactivation by evicting CTCF(CCCTC-binding factor)from the Xist(X inactive specific transcript)promoter,which enables Xist expression in mouse embryonic stem cells(ESCs)87,88.(C)MEG3(maternally expressed gene 3)recruits PRC2(polycomb repressive complex 2)to TGF-target genes via GA-rich sequences to form an RNADNA triplex,resulting in suppressed breast cancer cell growth and invasion89.(D)NEAT1(nuclear paraspeckle assembly transcript 1)is the architectural lncRNA of paraspeckles 90 that regulate multiple cellular progresses including mitochondrialstress(mito-stressors)75.NEAT1 has two isoforms:the 3 polyadenylated NEAT1_1 and the 3 triple helix ended NEAT1_2.(E)MALAT1(metastasis-associated lung adeno-carcinoma transcript 1)is localized at nuclear speckles 36 and sequesters phosphorylated SR(serine and arginine-rich)proteins to regulate pre-mRNA splicing 37,9193.(F)Sno-lncRNAs(lncRNA with snoRNA ends)94 and SPAs(5 snoRNA capped and 3 polyadenylated lncRNA)95 derived from the minimal micro-deletion associated withPraderWilli syndrome accumulate at their sites of transcription and interact with RNA-binding proteins to modulate mRNA alternative splicing.(G)SLERT(snoRNA-endedlncRNA that enhances pre-rRNA transcription)is exclusively localized in DFCs(dense fibrillar components)in the human nucleolus 96 and promotes pre-rRNA transcriptionin nucleoli 40.(H)Schematic depiction of lncRNAs(linc-RoR 77 and lncRNA-PNUTS 78)that function as miRNA sponges in cytosol.(I)Schematic of lncRNAs,such asMACC1-AS1 79 and 1/2 sbsRNAs 97,that regulate mRNA turnover in the cytosol.(J)ZFAS1(ZNFX1 antisense RNA 1)80,H19(H19 imprinted maternally expressed tran-script)81,and TUG1(taurine upregulated 1)81 are involved in ribosome production and assembly.(K)RMRP(RNA component of mitochondrial RNA-processingendoribonuclease)82 and SAMMSON(survival-associated mitochondrial melanoma-specific oncogenic non-coding RNA)83,84 are localized to mitochondria toregulate mi-tochondrialhomeostasisandmetabolism.Abbreviations:Chr.,chromosome;FC,fibrillarcenter;GC,granularcomponent;PSP,paraspeckleprotein;RBP,RNA-bindingprotein.Trends in Biochemical SciencesAn official publication of the INTERNATIONAL UNION OF BIOCHEMISTRY ANDMOLECULAR BIOLOGYTrends in Biochemical Sciences,November 2020,Vol.45,No.11949splicing efficiencies than nuclear RNAs 19 and that lncRNAs in general are less efficiently splicedthanmRNAs4,14.Theinefficientsplicing oflncRNAsisunlikelytoreflect genearchitecture,geno-mic position,or expression levels 19,and is at least in part due to weak internal splicing signalssuchasthepolypyrimidinetractandthedistancebetween3splicesiteandbranchpointinlncRNAgenes 19.Consistently,binding of U2AF65,which determines the splice site and facilitatesRNA splicing 26,was depleted in lncRNAs 4.Further,Pol II occupancy also appears tocorrelate with lncRNA splicing and nuclear localization,as indicated by observations thatincreased Pol II occupancy on introns correlated with reduced lncRNA splicing efficiencyand that decreased Pol II occupancy on promoter-proximal regions was associated withlncRNA nuclear enrichment 19.Such a correlation between inefficient lncRNA splicingTrendsTrends inin BiochemicalBiochemical Sciences SciencesFigure 2.Distinct Transcription and Processing of Long Non-Coding RNAs(lncRNAs)Leads to NuclearRetention.(A)Schematic of mRNA transcription,processing,and translation.Promoters of mRNA genes exhibit multiplehistone modifications and recruit TFs(transcription factors).The vast majority of mRNAs are transcribed by Pol II(RNApolymerase II)(CTD S5P,serine 5 phosphorylation of the C-terminal domain),are efficiently spliced 14,14,19,20,andare exported to the cytoplasm by the TREX(transcription export)complex where they are translated to generate functionalproteins 17.(B)Schematic of lncRNA transcription,processing,and function.The promoters of lncRNA genes usuallyexhibit histone H3 lysine 9 trimethylation(H3K9me3)and recruit fewer TFs.Many lncRNAs are transcribed by dysregulatedPol II(CTD S2P and S5P),have 23 exons,and are inefficiently spliced and polyadenylated 14,14,19,20;some aredegraded on chromatin 20.Many lncRNAs are retained in the nucleus(Figure 1AG)whereas others are exported to thecytoplasm by NXF1(nuclear RNA export factor 1)17 to modulate cellular processes(Figure 1HK).(C)A number ofantisense lncRNAs form R-loop structures when transcribed.The formation of an R-loop structure of Knps1(antisenseorientation to SPHK1)32,ANRIL(antisense non-coding RNA in the INK4 locus)33,and VIM-AS1(antisense of vimentin)34 allows these transcripts to be restricted to the nucleus.Abbreviations:RBP,RNA-binding protein;lncRNP,lncRNAprotein complex.Trends in Biochemical SciencesAn official publication of the INTERNATIONAL UNION OF BIOCHEMISTRY ANDMOLECULAR BIOLOGY950Trends in Biochemical Sciences,November 2020,Vol.45,No.11and increased nuclear retention 19 is reminiscent of mRNA processing because splicingpromotes mRNA recruitment to the export machinery,and functional coupling betweensplicing and mRNA export is a conserved and general fe