Regulation of gene expression bycis- acting.pdf

High-throughput sequencing technologies and chroma-tin state maps have shown that eukaryotic cells produce a plethora of non-coding transcripts13.Of these,long non-coding RNAs(lncRNAs)are defined arbitrarily as transcripts of more than 200 nucleotides that do not belong to any other well-defined group of non-coding RNAs,such as ribosomal RNAs.Through various mecha nisms,lncRNAs have been implicated in a wide array of cellular processes,including transcriptional regulation,differentiation,cellular reprogramming and many others(reviewed elsewhere46).With varying levels of evidence,lncRNAs have also been implicated in vari-ous human diseases79.lncRNAs are transcribed by RNA polymerase II(Pol II),and their biogenesis is similar to mRNAs in that they are capped and polyadenylated.lncRNAs are also usually spliced,although their exon number and splicing efficiency are on average lower than those of mRNAs1013.However,as lncRNAs are predominantly defined by exclusion criteria,the set of genes annotated as lncRNAs includes many distinct sub-groups,exemplifying diverse structural and,presumably,functional characteristics.Assigning lncRNAs to distinct functional groups is essential to identify common prin-ciples,and thus comprises a pivotal step when beginning to elucidate their roles.This step remains very challeng-ing,with limited progress being made in the past decade of lncRNA research.One type of lncRNA classification is based on the location at which the lncRNA functions relative to its transcription site.Trans-acting lncRNAs are transcribed,processed and then vacate their sites of transcription to exert their function elsewhere,akin to mRNAs.Their final destination,be it in the cytoplasm or nucleus,does not depend on their transcription site.Accordingly,as long as their levels are properly maintained,transcrib-ing these lncRNAs from a different genomic location or supplanting them into the system should not interfere with their function(that is,their loss of function can be rescued by their expression from exogenous locations).A few examples of such lncRNAs have now been exten-sively characterized1416,and many additional lncRNAs have been ascribed trans activities1720.By contrast,cis-acting lncRNAs are those whose activity is based at and dependent on the loci from which they are transcribed.Transcripts with the poten-tial of acting in cis likely make up a substantial portion of known lncRNAs:the majority of lncRNAs are enriched in the chromatin fraction,and specifically are tethered to chromatin presumably at their sites of transcription through Pol II21.If functional,this would indicate that the effects of these lncRNAs are centred at these loci.In addition,the fairly low levels at which lncRNAs are generally expressed,oftentimes just a few molecules per cell13,22,23,naturally favour a cis mechanism of action,as diffusion or transport to other cellular compartments would render these transcripts too diluted to mediate a plausible function.Functions in cis regulation are also consistent with the evolutionary trajectories of lncRNAs(Box1),which often lack substantial sequence conser-vation,yet are sometimes found in conserved locations relative to other genes across distant species24.Advancements in our understanding of chroma-tin organization and transcription regulation,and the development of perturbation techniques relevant to Regulation of gene expression by cis-acting long non-coding RNAsNoaGil and IgorUlitsky *Abstract|Long non-coding RNAs(lncRNAs)are diverse transcription products emanating from thousands of loci in mammalian genomes.Cis-acting lncRNAs,which constitute a substantial fraction of lncRNAs with an attributed function,regulate gene expression in a manner dependent on the location of their own sites of transcription,at varying distances from their targets in the linear genome.Through various mechanisms,cis-acting lncRNAs have been demonstrated to activate,repress or otherwise modulate the expression of target genes.We discuss the activities that have been ascribed to cis-acting lncRNAs,the evidence and hypotheses regarding their modes of action,and the methodological advances that enable their identification and characterization.The emerging principles highlight lncRNAs as transcriptional units highly adept at contributing to gene regulatory networks and to the generation of fine-tuned spatial and temporal gene expression programmes.Department of Biological Regulation,Weizmann Institute of Science,Rehovot,Israel.*e-mail:igor.ulitskyweizmann.ac.ilhttps:/doi.org/10.1038/s41576-019-0184-5REVIEWSNature reviews|Geneticsthe study of lncRNAs(most notably CRISPR-related technologies),have facilitated the in-depth character-ization of several cis-acting lncRNAs that operate at various genomic distances and through apparently distinct mechanisms of action.From these studies,a broader picture of the roles of cis-acting lncRNAs is starting to emerge.In this Review,we begin with an overview of the prevalence,genomic locations,orien-tations and conservation of cis-acting lncRNAs,as well as the functional clues provided by these attributes.We describe the types of cis effects that lncRNAs can have on their targets,supplementing with recent examples of fairly comprehensively studied lncRNAs.Finally,we elaborate on reported and possible mechanisms of action employed by cis-acting lncRNAs,as well as the tools currently being used or developed for investigating their functions.Discerning between trans-and cis-acting lncRNAsAlthough the associations of a lncRNA with its site of transcription,genomic neighbourhood,epigenetic envi-ronment and mode of conservation(that is,sequence versus synteny)can all point to a lncRNA acting in cis,they do not provide direct proof.For example,some lncRNAs co-expressed with their neighbouring gene seem to have no cis-regulatory activity but rather coop-erate with the product of this gene20.Some experimental indication for cis activity can be achieved by express-ing the lncRNA from a different genomic location or supplanting it into the system,for example through transfection of a vector.If such trans supplementation does not rescue the phenotype,the lncRNA is suspected of working in cis.It is noteworthy that some lncRNAs that act near their transcription sites can be rescued by trans expression,indicating they are not true cis-acting lncRNAs.For example,the FIRRE lncRNA,which orchestrates spatial proximity between loci found on different chromosomes,operates from its site of tran-scription25,although expressing the lncRNA in trans can rescue a FIRRE-null phenotype26.Similarly,the NEAT1 lncRNA serves as a scaffold for the assembly of para-speckles,nuclear membrane-less organelles that form near its transcription site27.Paraspeckle-like foci can also be formed when NEAT1 is expressed exogenously28,although their functionality has not yet been tested.Nonetheless,the success of trans supplementation is highly dependent on exact timing and expression levels,and so failure in such a setting does not prove cis activity.Additionally,even if trans supplementation does work,it does not unequivocally prove that the lncRNA does not work in cis,as it is theoretically possible that when the lncRNA floods the system,a sufficient amount reaches the vicinity of its site of transcription.More direct experimental proof can be gained using cells or tissues generated from a mixed genetic background,such as through crosses between mouse strains2933.This approach allows perturbation of only a single lncRNA allele and then exploitation of single-nucleotide poly morphisms to discern whether the effect is limited to target genes located on the same allele as the per-turbed lncRNA,which would confirm a cis mechanism of action.Box 1|evolution of cis-acting long non-coding RnAsturnover of long non-coding rNa(lncrNa)loci is a major component of genome evolution.as discussed extensively elsewhere24,emergence of new lncrNas genes generally requires a combination of a promoter and polyadenylation sequences as well as potentially some splicing signals.accumulation of the correct combination of mutations that is,mutations that enable transcription initiation,recruitment of rNa processing factors and suppression of transcription termination might lead to lncrNa production in a previously transcriptionally silent locus.this lncrNa,or the act of its transcription,might then acquire cis functions(see the figure,part a;yellow structure denotes lncrNa transcript).as enhancers already contain elements for the recruitment of rNa polymerase ii(Pol ii),it might be possible that few sequence changes are sufficient to gain or lose lncrNa production at enhancers(see the figure,part b;yellow wavy lines denote enhancer rNas).indeed,one study identified hundreds of regulatory regions with repurposed activities among closely-related mammalian species,that is,regions with promoter characteristics in one species and enhancer characteristics in another160.interestingly,most of these regions likely served as enhancers in the ancestral species,which then acquired species-specific promoter abilities,demonstrating that enhancers can serve as fertile ground for the creation of novel transcripts.it is tempting to speculate that enhancers which acquire promoter capabilities do not necessarily or immediately lose their enhancer activity.rather,other features of the enhancer locus(for example,spatial proximity to a target gene)can favour the creation of a transcript which participates in the same circuit as the enhancer.another important,albeit rare,source of new functional lncrNas is disruption of protein-coding gene(PCG)open reading frames(OrFs)161,162(see the figure,part c;green wavy lines denote mrNas).this process may affect preferentially PCGs that have paralogues in the genome,so that the original PCG function is maintained162.One could hypothesize that,compared with enhancer-derived lncrNas,PCG-derived lncrNas are more likely to act in trans,as the sequences they inherit from their PCG ancestors would facilitate higher(and perhaps more cytoplasmic)expression162.Nonetheless,such lncrNas may hold on to cis functionality,as some PCG loci seem to induce cis effects similar to lncrNa loci.Over time,a cis-acting rNa can accumulate sequences that will facilitate trans-acting activities and vice versa(see the figure,part d;coloured circles denote trans factors).as most lncrNas are found in both the cytoplasm and the nucleus163,164,it is possible that some lncrNas carry both functions,although such examples remain scarce107.MutationTranspositionGenome rearrangementSpecies ASpecies BabEnhancerlncRNAEnhancerPCGEnhancerEnhancerEnhancerPCGEnhancerEnhancerEnhancercdPCGlncRNA locations of cis-acting lncRNAsIf functional,lncRNAs that are enriched around their own sites of transcription are expected to participate in chromatin-related processes,such as the modula-tion of chromatin structure,chromatin modifications or transcription control.The contribution of lncRNAs to these processes in cis is evidenced by the enrichment of lncRNA genes in the vicinity of regulatory sequences of the genome.Most notably,various studies have indicated that 3060%of lncRNAs are transcribed from regions showing characteristics of enhancers21,3436,with the vari-ability arising from the different methodologies and parameters used for annotating both lncRNAs and enhancers.When considering the commonly used annotations of enhancers by chromHMM37,lncRNA transcription start sites(TSSs)tend to overlap regions classified as enhancers more than protein-coding gene(PCG)TSSs do(Fig.1a).This observation is in agreement with findings showing that many lncRNA promoters exhibit DNA motifs and protein-binding characteristics associated with both promoters and enhancers36,38.An additional indication for the widespread cis-regulatory function of lncRNAs is their enrichment in the vicinity of genes belonging to specific categories.Various EnhancersDNA regulatory elements that activate gene transcription.Enhancers can operate from nearby or within their targets,or across large genomic distances.bcad*Translation regulatorStructural moleculeAntioxidantBindingCatalysisMolecular function regulatorTransporterMolecular transducerCargo receptor050,000100,000150,000200,000250,000300,000Distance(bp)NoneSequenceSequence andsyntenySyntenyMedian number of lncRNAs within 1MbTotalConservedPCGslncRNAs0510152025Transcription regulatorTranslation regulatorStructural moleculeAntioxidantBindingCatalysisMolecular function regulatorTransporterMolecular transducerCargo receptorTSS stateEnhancer stateFraction of epigenomes0.00.20.40.60.81.0Transcription regulatorFig.1|Genomic locations and prevalence of cis-acting long non-coding RnAs.a|The fraction of transcription start sites(TSSs)of RefSeq protein-coding gene(PCG)transcripts or GENCODE v30 long non-coding RNA(lncRNA)transcripts that overlap regions classified as TSSs or,conversely,as enhancers in a 15-state chromHMM model of 127 distinct epigenomes165,166.Asterisks indicate P 2.21016,calculated using the two-sided Wilcoxon test.b|The distribution of the distances between the TSSs of human PCGs belonging to the indicated categories of gene ontology(GO)terms and the closest lncRNA transcript.Shown are all GO term categories under molecular function that contain at least 50 genes.Highlighted in pink is the GO term category transcription regulator activity.c|The median number of lncRNA transcripts found within 1 Mb of the TSS of PCGs that belong to the indicated categories of GO terms(pink:the median number of sequence-conserved lncRNAs within that same distance).Shown are all GO term categories under molecular function that contain at least 50 genes.d|The conservation type for the lncRNA found closest to TSSs of PCGs belonging to the transcription regulator activity GO category.Nature reviews|GeneticsReviewsanalyses have shown that the protein products of genes that flank lncRNA genes are involved in transcription-related processes1,23,3941.A similar analysis of more recent lncRNA annotations shows that genes involved in processes related to transcription regulation,such as transcription factors(TFs)or chromatin remodellers,tend to have a closer lncRNA neighbour than do other genes(Fig.1b).Genes involved in transcription regulation also tend to be surrounded by more lncRNA genes,suggesting that multiple lncRNAs might cooperate in their regula-tion;this trend is more pronounced when considering the number of conserved lncRNAs41(Fig.1c).Importantly,although somewhat weaker,this trend is maintained when normalizing to the number of PCGs within the same genomic distance,or to the number of annotated enhancer elements(not shown),indicating that this enrichment does not reflect merely an increase in regu-latory intergenic space.The nearest lncRNA neighbours of transcription-related genes often have conserved positions rather than sequences(Fig.1d),pointing to the potential im