Noncoding RNAs play surprising roles in regulating transcription
Finally, we describe a newly recognized, unexpected mode of transcriptional regulation. As already discussed in Chapter 10, one of the biggest surprises that has emerged from studying transcription at the genome-wide level is the pervasive transcriptional activity throughout the genome. As a recent review puts it, “transcription is achieved using a breathtaking number of transcription events.” Approximately 180,000 mouse cDNAs have been identified, whereas a mere ~20,200 protein-coding mouse genes are known. Less than 3% of the human genome encodes proteins, leaving many of the remaining transcribed RNA sequences with unknown functions. Analysis of the mouse and human transcriptomes thus reveals a huge number of transcripts that do not encode proteins, hence their name noncoding RNAs or ncRNAs. A major question still lingers in the transcription field: are these ncRNAs mere transcriptional noise, resulting from initiation at spurious promoters that have arisen serendipitously in the genome, or do some of them have bona fide cellular functions? If yes, what are these mysterious functions whose existence we did not suspect until now? This is the dark matter of molecular biology. There is another possibility: the functional importance of ncRNA transcription may lie in the process of transcription itself rather than in the products. Such views presume that the changes occurring in chromatin structure during transcription are what matters, since transcribed regions are seemingly more open and this openness facilitates further rounds of transcription. Thus, it is possible that transcription of ncRNAs helps to maintain a transcriptionally competent state for the transcription of functional RNA- and protein-coding genes. In the few cases where we have information, the function of these molecules seems remarkably diverse (Figure 1).
Fig1. Mechanisms of lncRNA mediated regulation of transcription in eukaryotes. (A) In response to stress, the long noncoding RNA (lncRNA) transcribed upstream of the Cyclin D1 gene interacts with the RNA-binding protein TLS, translocated in liposarcoma, and induces a conformational change in the protein. The activated TLS inhibits histone acetyltransferase activities and thus inhibits Cyclin D1 transcription. (B) The promoter of the DHFR (dihydrofolate reductase) gene is inhibited by binding of the lncRNA to the basal transcription factor TFIIB, preventing initiation. The mechanism also involves the formation of a stable triplex com plex, purine-purine-pyrimidine, between the single-stranded lncRNA and the double-stranded DHFR promoter. (C) The lncRNA Evf1-Evf2 interacts with the homeodomain protein Dlx2 to activate the enhancer serving Dlx5 and Dlx6, two genes that are involved in neuronal differentiation and migration. Expression of Evf2 is highly regulated in the developing mouse brain. (D) Epigenetic silencing of gene clusters in cis by lncRNAs: the example shown is Xist, the lncRNA crucial for X-chromosome inactivation in mammals. Xist coats the X-chromosome that is to be stably inactivated. It is believed that Xist coating establishes a specialized nuclear compartment devoid of Pol II; it also interacts with EZH2, the subunit of PRC2 that introduces repressive methylation marks on H3K27. (E) Epigenetic repression of genes in trans. The HOTAIR lncRNA is transcribed within the human HOXC gene cluster; it then directly targets epigenetic modifiers, such as PRC2, to the HOXD gene cluster. Other histone-modifying complexes, such as G9a methyltransferase, may also be targeted.
The sizes and genomic locations of noncoding transcripts are remarkably diverse
Noncoding RNAs may be relatively short or relatively long, with the arbitrary dividing line set at ~200 nucleotides. There are several well-defined classes of short ncRNAs; we describe their biogenesis in Chapter 14 and the roles of others in Chapter 17. Here we discuss the properties and transcription regulatory functions of some long non coding RNAs or lncRNAs, all recently described.
Many ncRNAs tend to be transcribed away from the 5′- or 3′-ends of protein-coding genes; for the lncRNAs, there is significant concentration near promoters, initial exons, and initial introns. At least some of these RNA molecules are both capped and polyadenylated, as are regular mRNA molecules. Some also contain short, stable stem–loop structures that are deemed important in protein binding. The proteins bound could be either basal or gene-specific transcription factors, elongation factors, and/or chromatin remodelers. Figure 1 presents several well understood lncRNAs that function as regulators of transcription. The mechanisms involved are incredibly varied and show no common feature. Figure 2 illustrates another fascinating example of how one lncRNA mediates the apoptotic response of cells to stress, by repressing the transcription of pro-survival genes.
Fig2. Participation of lincRNA in p53-mediated transcriptional repression. p53 is an important tumor-suppressor gene mutated in around 50% of human cancers. The p53 protein binds to DNA as a tetramer; it becomes stabilized in response to DNA damage and triggers a complex transcriptional response, activating or repressing the expression of numerous genes. The transcriptional response leads either to cell-cycle arrest or to apoptosis (programmed cell death). The p53-mediated gene repression that is part of apoptosis occurs through a mechanism involving a long intergenic noncoding RNA (lincRNA-p21), so named for its physical proximity to the p53 target gene p21, although p21 is not involved in the lincRNA pathway. The sequence of events is as follows: p53 induces the transcription of lincRNA-p21, which then interacts with a repressive RNA–protein complex, hnRNP-K; the interaction is required for proper localization of hnRNP-K onto responsive genes. The repressor complex contains the linker histone variant H1.2 and blocks histone acetylation by the histone acetyltransferase p300. The changes in chromatin organization are the probable cause for transcriptional repression of the linked pro-survival genes, leading to cell death.
Probably the best-studied example of lncRNAs is the Xist RNA that participates in X-chromosome inactivation (XCI). It has been known for years that somatic cells of female mammals contain two X chromosomes, whereas male cells have one X and one Y sex chromosome. To prevent gross imbalance between females and males in the expression of genes present on the X chromosome, nature has come up with a solution. During early embryonic development of female embryos, one of the X chromosomes becomes fully inactivated in a complex sequence of molecular events, collectively known as X-chromosome inactivation, XCI. Once the inactivation has taken place, the chromosome remains silent for the rest of the life of that individual.
Only recently are we starting to understand the molecular events that lead to XCI. For about 25 years, all attention was focused on a lncRNA, called Xist, but the mechanisms of its action remained elusive. Only a combination of several state-of-the-art methods, both in vitro and in vivo, allowed a clearer picture to emerge. Several screening procedures allowed the identification of a protein, called SPEN, that binds Xist during early stages of inactivation. SPEN is immediately recruited to the X chromosome upon the up-regulation of the gene that encodes Xist and is targeted to enhancers and promoters of active genes. SPEN is a very large protein (~400 kDa) that contains four RNA recognition motifs (RRMs; see Chapter 6), a nuclear receptor interaction domain, and a SPOC domain. Of interest, it was found that only RRM domains 2 to 4 and the SPOC domain were responsible for SPEN accumulation on the inactive X chromosome, as well as for X-linked gene silencing. Although SPEN coats the X chromosome immediately upon Xist induction, it contacts chromatin only at active promoters and enhancers, which are the substrates for SPEN-mediated gene silencing. SPEN association with chromatin is favored by active transcription, as SPEN disengages from chromatin when X-linked genes become silenced. A simplified schematic of the process is presented in Figure 3. SPEN bridges Xist to the transcriptional machinery, histone deacetylases, and remodeling factors to ensure robust and efficient XCI.
Fig3. Mechanism of SPEN action in X-chromosome inactivation. The process of inactivation is initiated by SPEN binding to active gene promoters and enhancers. The recognition of active promoters involves interaction with RNAP II. SPEN also recruits the enzyme HDA3 (histone deacetylase 3) that removes gene-activating acetyl groups from histones. In addition, SPEN is associated with components of the nucleosome remodeling and deacetylase (NuRD) complex, another epigenetic silencer.
Finally, lncRNAs have been implicated in cell cycle progression. In a 2020 study, published in Cell Reports, nascent RNAs were isolated from a human cell line; more than 900 lncRNAs were identified whose synthesis peaks during the S phase. The role of three of these lncRNAs was assessed in knockdown experiments; all three were required for normal S phase progression, acting on distinct regulatory networks. The large number of these RNAs, each expressed at a specific time period during the S phase, emphasized the importance of gene expression control exerted by lncRNAs.
The diversity of mechanisms is amazing. One thing is certain: we have seen just the tip of the iceberg as far as lncRNAs and their functions are concerned. More insights were gleaned from the ENCODE project, which systematically analyzed the entire human genome and transcriptome.
